refactor(context): trim verbose comments

A configured tool_output_max_bytes smaller than the truncation notice
itself can't fit a bounded preview, so a persisted result could exceed the
ceiling. Enforce a config floor (ge=1024) so nonsensical values are
rejected at load time instead of being worked around at runtime.

.github/workflows/build-release.yml

@@ -21,8 +21,6 @@ jobs:
             target: macos-x86_64
           - os: ubuntu-22.04
             target: linux-x86_64
-          - os: ubuntu-22.04-arm
-            target: linux-arm64
           - os: windows-latest
             target: windows-x86_64
 
@@ -45,20 +43,6 @@ jobs:
           uv sync --frozen
           uv run pyinstaller strix.spec --noconfirm
 
-          if [[ "${{ runner.os }}" == "Windows" ]]; then
-            dist/strix.exe --version
-          else
-            dist/strix --version
-          fi
-
-          if [[ "${{ matrix.target }}" == "linux-arm64" ]]; then
-            file dist/strix
-            file dist/strix | grep -q "ARM aarch64" || {
-              echo "::error::linux-arm64 artifact is not an ARM aarch64 binary"
-              exit 1
-            }
-          fi
-
           VERSION=$(grep '^version' pyproject.toml | head -1 | sed 's/.*"\(.*\)"/\1/')
           mkdir -p dist/release
 

.gitignore

@@ -1,8 +1,8 @@
 # Node / local-viewer SPA source (the built bundle in
-# strix/interface/viewer/static/ is committed and shipped; do not ignore it)
+# strix/viewer/static/ is committed and shipped; do not ignore it)
 node_modules/
-strix/interface/viewer/frontend/node_modules/
-strix/interface/viewer/frontend/.vite/
+strix/viewer/frontend/node_modules/
+strix/viewer/frontend/.vite/
 
 # Python
 __pycache__/

CONTRIBUTING.md

@@ -102,16 +102,16 @@ We welcome feature ideas! Please:
 ## 🖥️ Local viewer SPA
 
 `strix view` serves a prebuilt web UI whose source lives in
-`strix/interface/viewer/frontend/` (a Vite + React project) and whose built output is
-committed to `strix/interface/viewer/static/` and shipped in the package. End users never
-run a JS build. If you change anything under `strix/interface/viewer/frontend/`, rebuild
+`strix/viewer/frontend/` (a Vite + React project) and whose built output is
+committed to `strix/viewer/static/` and shipped in the package. End users never
+run a JS build. If you change anything under `strix/viewer/frontend/`, rebuild
 and commit the output:
 
 ```bash
-make viewer   # or: cd strix/interface/viewer/frontend && npm ci && npm run build
+make viewer   # or: cd strix/viewer/frontend && npm ci && npm run build
 ```
 
-Commit both the source change and the regenerated `strix/interface/viewer/static/`.
+Commit both the source change and the regenerated `strix/viewer/static/`.
 
 ## 🤝 Community
 

Makefile

@@ -69,8 +69,8 @@ clean:
 
 viewer:
 	@echo "🖥️  Building the local-viewer SPA..."
-	cd strix/interface/viewer/frontend && npm ci && npm run build
-	@echo "✅ Viewer built to strix/interface/viewer/static/ (commit the changes)."
+	cd strix/viewer/frontend && npm ci && npm run build
+	@echo "✅ Viewer built to strix/viewer/static/ (commit the changes)."
 
 dev: format lint type-check
 	@echo "✅ Development cycle complete!"

docs/usage/cli.mdx

@@ -61,28 +61,11 @@ strix (--target <target> | --target-list <path> | --mount <path>) [options]
   Path to a custom config file (JSON) to use instead of `~/.strix/cli-config.json`.
 </ParamField>
 
-<ParamField path="--max-budget" type="number">
+<ParamField path="--max-budget-usd" type="number">
   Maximum LLM spend in USD for the whole scan, counted cumulatively across the
   root agent and every child agent. The budget is checked after each model
-  response.
-
-  In non-interactive mode (`-n`), once the running cost reaches the threshold,
-  the scan stops cleanly with a `stopped` status (not a failure) and the sandbox
-  is torn down. Sub-agents are stopped early, at 90% of the budget, reserving
-  the final slice for the root agent to wind down and produce the final report.
-
-  In interactive mode, reaching the budget pauses the scan instead of ending
-  it: every agent parks, and sending any message resumes the scan with the cap
-  extended by the original budget amount. There is no sub-agent reserve in
-  interactive mode.
-
-  As the budget is approached, graduated wrap-up warnings are surfaced to
-  **every** agent so they can finish their work and call their lifecycle tool
-  before the hard stop. The bands sit just below each role's own stop point: the
-  root is warned at **70%, 85% and 95%** (it stops at 100%), while sub-agents are
-  warned at **75%, 80% and 85%** (they stop at the 90% reserve). In interactive
-  mode every agent uses the **70%, 85% and 95%** bands. Percentages shown in the
-  warnings are the real cumulative spend against the full budget.
+  response; once the running cost reaches the threshold, the scan stops cleanly
+  with a `stopped` status (not a failure) and the sandbox is torn down.
 
   Must be greater than `0`. Omit the flag for no limit.
 
@@ -101,19 +84,6 @@ strix (--target <target> | --target-list <path> | --mount <path>) [options]
     counts.
 </ParamField>
 
-<ParamField path="--max-turns" type="integer" default="500">
-  Maximum number of turns (one model response plus its tool round) allotted to
-  **each** agent, applied per run. When an agent reaches this limit it is
-  force-stopped.
-
-  As the limit is approached, graduated wrap-up warnings (at 70%, 85% and 95%)
-  are injected into that agent's next model turn so it can prioritise its
-  remaining work and call its lifecycle tool (`finish_scan` for the root agent,
-  `agent_finish` for sub-agents) before the hard stop.
-
-  Must be greater than `0`.
-</ParamField>
-
 ## Examples
 
 ```bash
@@ -129,9 +99,6 @@ strix --target api.example.com --instruction "Focus on IDOR and auth bypass"
 # CI/CD mode
 strix -n --target ./ --scan-mode quick
 
-# Cap cost and per-agent turns
-strix --target https://example.com --max-budget 25 --max-turns 300
-
 # Force diff-scope against a specific base ref
 strix -n --target ./ --scan-mode quick --scope-mode diff --diff-base origin/main
 

pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "strix-agent"
-version = "1.4.0"
+version = "1.3.1"
 description = "Open-source AI Hackers for your apps"
 readme = "README.md"
 license = "Apache-2.0"
@@ -79,10 +79,10 @@ build-backend = "hatchling.build"
 
 [tool.hatch.build.targets.wheel]
 packages = ["strix"]
-# The prebuilt viewer bundle under strix/interface/viewer/static/ ships automatically
+# The prebuilt viewer bundle under strix/viewer/static/ ships automatically
 # (hatchling includes non-.py files under the package). The Vite SOURCE lives
-# under the package dir too (strix/interface/viewer/frontend/) but must never ship in the wheel.
-exclude = ["strix/interface/viewer/frontend", "strix/interface/viewer/frontend/**"]
+# under the package dir too (strix/viewer/frontend/) but must never ship in the wheel.
+exclude = ["strix/viewer/frontend", "strix/viewer/frontend/**"]
 
 # ============================================================================
 # Type Checking Configuration
@@ -222,10 +222,10 @@ ignore = [
 "tests/test_codex_streaming.py" = ["N802"]
 "tests/test_report_pdf.py" = ["S105", "S106"]
 # Stdlib HTTP handler overrides (do_GET/do_POST) and lazy imports that avoid a
-# circular dependency with strix.telemetry / strix.interface.viewer.report_pdf.
-"strix/interface/viewer/server.py" = ["N802", "PLC0415"]
+# circular dependency with strix.telemetry / strix.viewer.report_pdf.
+"strix/viewer/server.py" = ["N802", "PLC0415"]
 # Lazy telemetry import to avoid importing PostHog before the viewer starts.
-"strix/interface/viewer/cli.py" = ["PLC0415"]
+"strix/viewer/cli.py" = ["PLC0415"]
 # Lazy imports inside functions to avoid circular dependency with
 # strix.telemetry / strix.report.dedupe / cvss.
 "strix/tools/notes/tools.py" = ["PLC0415", "TC002"]

scripts/install.sh

@@ -41,7 +41,7 @@ fi
 
 combo="$os-$arch"
 case "$combo" in
-  linux-x86_64|linux-arm64|macos-x86_64|macos-arm64|windows-x86_64)
+  linux-x86_64|macos-x86_64|macos-arm64|windows-x86_64)
     ;;
   *)
     echo -e "${RED}Unsupported OS/Arch: $os/$arch${NC}"

strix.spec

@@ -26,7 +26,7 @@ for tcss_file in strix_root.rglob('*.tcss'):
     datas.append((str(tcss_file), str(rel_path.parent)))
 
 # Prebuilt local-viewer SPA (served by `strix view`).
-viewer_static = strix_root / 'interface' / 'viewer' / 'static'
+viewer_static = strix_root / 'viewer' / 'static'
 for asset in viewer_static.rglob('*'):
     if asset.is_file():
         rel_path = asset.relative_to(project_root)
@@ -158,12 +158,12 @@ hiddenimports = [
     'strix.report.dedupe',
     'strix.report.state',
     'strix.report.writer',
-    'strix.interface.viewer',
-    'strix.interface.viewer.auth',
-    'strix.interface.viewer.cli',
-    'strix.interface.viewer.report_pdf',
-    'strix.interface.viewer.server',
-    'strix.interface.viewer.transcript',
+    'strix.viewer',
+    'strix.viewer.auth',
+    'strix.viewer.cli',
+    'strix.viewer.report_pdf',
+    'strix.viewer.server',
+    'strix.viewer.transcript',
 
     # PDF report generation + encryption
     'reportlab',

strix/agents/factory.py

@@ -34,7 +34,7 @@ from strix.tools.notes.tools import (
     list_notes,
     update_note,
 )
-from strix.tools.output_store import bound_and_store, bound_text
+from strix.tools.output_store import bound_text
 from strix.tools.proxy.tools import (
     list_requests,
     list_sitemap,
@@ -43,12 +43,7 @@ from strix.tools.proxy.tools import (
     view_request,
     view_sitemap_entry,
 )
-from strix.tools.reporting.tool import (
-    create_dependency_report,
-    create_vulnerability_report,
-    get_report,
-    list_reports,
-)
+from strix.tools.reporting.tool import create_dependency_report, create_vulnerability_report
 from strix.tools.thinking.tool import think
 from strix.tools.todo.tools import (
     create_todo,
@@ -115,11 +110,11 @@ def _tool_output_limits() -> tuple[int, int]:
     return context.tool_output_max_lines, context.tool_output_max_bytes
 
 
-async def _bound_result(result: Any) -> Any:
+def _bound_result(result: Any) -> Any:
     if not isinstance(result, str):
         return result
     max_lines, max_bytes = _tool_output_limits()
-    return await bound_and_store(result, max_lines=max_lines, max_bytes=max_bytes)
+    return bound_text(result, max_lines=max_lines, max_bytes=max_bytes)
 
 
 def _format_tool_error(exc: Exception) -> str:
@@ -135,7 +130,7 @@ def _with_bounded_result(tool: FunctionTool) -> FunctionTool:
     invoke_tool = tool.on_invoke_tool
 
     async def invoke(ctx: Any, raw_input: str) -> Any:
-        return await _bound_result(await invoke_tool(ctx, raw_input))
+        return _bound_result(await invoke_tool(ctx, raw_input))
 
     tool.on_invoke_tool = invoke
     tool._strix_bounded = True  # type: ignore[attr-defined]
@@ -147,7 +142,7 @@ def _function_tool_with_error_result(tool: FunctionTool) -> FunctionTool:
 
     async def invoke(ctx: Any, raw_input: str) -> Any:
         try:
-            return await _bound_result(await invoke_tool(ctx, raw_input))
+            return _bound_result(await invoke_tool(ctx, raw_input))
         except Exception as exc:  # noqa: BLE001 - tool errors should be model-visible results.
             logger.debug("Tool %s failed; returning error as result", tool.name, exc_info=True)
             return _format_tool_error(exc)
@@ -162,7 +157,7 @@ def _custom_tool_as_function_tool(tool: CustomTool) -> FunctionTool:
         if not custom_input:
             return f"`{_custom_tool_input_field(tool)}` must be a non-empty string."
         try:
-            return await _bound_result(await tool.on_invoke_tool(ctx, custom_input))
+            return _bound_result(await tool.on_invoke_tool(ctx, custom_input))
         except Exception as exc:  # noqa: BLE001 - matches SDK CustomTool error-as-result behavior.
             logger.debug("Tool %s failed; returning error as result", tool.name, exc_info=True)
             return _format_tool_error(exc)
@@ -199,7 +194,7 @@ def _bound_custom_tool(tool: CustomTool) -> CustomTool:
     invoke_tool = tool.on_invoke_tool
 
     async def invoke(ctx: Any, raw_input: str) -> Any:
-        return await _bound_result(await invoke_tool(ctx, raw_input))
+        return _bound_result(await invoke_tool(ctx, raw_input))
 
     tool.on_invoke_tool = invoke
     return tool
@@ -415,8 +410,6 @@ _BASE_TOOLS: tuple[Tool, ...] = (
     web_search,
     create_vulnerability_report,
     create_dependency_report,
-    list_reports,
-    get_report,
     list_requests,
     view_request,
     repeat_request,

strix/agents/prompts/system_prompt.jinja

@@ -188,7 +188,7 @@ EFFICIENCY TACTICS:
   script fail with `ModuleNotFoundError`.
 - `exec_command` runs each command in a fresh non-interactive shell (plain
   pipes, no TTY). To drive an interactive or long-running process with
-  `write_stdin` — REPLs, `ssh`/`nc`/`ftp`, `sqlmap`, or to send Ctrl-C —
+  `write_stdin` — REPLs, `ssh`/`nc`/`ftp`, `msfconsole`, or to send Ctrl-C —
   you MUST start it with `exec_command(cmd="...", tty=true)` and then
   `write_stdin(session_id=<id>, chars="...")`. Calling `write_stdin` on a
   default (non-TTY) command or on a process that has already exited fails with
@@ -215,7 +215,6 @@ VALIDATION REQUIREMENTS:
 - A vulnerability is ONLY considered reported when a reporting agent uses create_vulnerability_report (or create_dependency_report for known-CVE dependency/supply-chain findings) with full details. Mentions in agent_finish, finish_scan, or generic messages are NOT sufficient
 - Reporting and fixing are ONE step, not two: when source is available, the reporting agent derives the concrete fix and files it INLINE via create_vulnerability_report (`code_locations` with `fix_before`/`fix_after` + `fix_pr_body`) — the report is not complete without it. Do NOT report first and then spawn a separate downstream agent to re-derive and re-apply the same patch; that just re-does the analysis and wastes tokens. (Do not silently patch a finding WITHOUT filing a report — the report, with its embedded fix, is the deliverable.)
 - DEDUPLICATION: The create_vulnerability_report tool uses LLM-based deduplication. If it rejects your report as a duplicate, DO NOT attempt to re-submit the same vulnerability. Accept the rejection and move on to testing other areas. The vulnerability has already been reported by another agent
-- REVIEWING FILED FINDINGS (orchestrator/root agent): use list_reports to see every vulnerability filed so far in this scan (by any agent, root or child) — metadata-first with per-severity counts — and get_report to read one finding in full by its id. These are read-only orchestration tools: the root agent uses them to track coverage, avoid dispatching work on already-covered ground, assemble the finish_scan executive summary, and reason about attack-chaining across confirmed findings. Leaf/specialist agents should NOT call them — just do your assigned testing and file findings. Each entry shows which agent filed it (agent_name), and your own entries are flagged by_you. list_notes/get_note do the same for notes.
 </execution_guidelines>
 
 <vulnerability_focus>

strix/config/models.py

@@ -429,45 +429,3 @@ def is_known_openai_bare_model(model_name: str) -> bool:
         return False
     entry = litellm.model_cost.get(name)
     return bool(entry and entry.get("litellm_provider") == "openai")
-
-
-def is_claude_model(model_name: str) -> bool:
-    return "claude" in (model_name or "").strip().lower()
-
-
-def is_bedrock_route(model_name: str) -> bool:
-    name = (model_name or "").strip().lower()
-    return name.startswith("bedrock/") or "anthropic." in name
-
-
-def _prompt_cache_name_candidates(model_name: str) -> list[str]:
-    # LiteLLM's model map keys the same model under several names; strip the
-    # route prefix, then leading dotted segments (region, provider).
-    name = (model_name or "").strip().lower()
-    for prefix in ("litellm/", "bedrock/"):
-        if name.startswith(prefix):
-            name = name[len(prefix) :]
-            break
-    candidates = [name]
-    rest = name
-    while "." in rest:
-        rest = rest.split(".", 1)[1]
-        candidates.append(rest)
-    return candidates
-
-
-def bedrock_route_supports_prompt_caching(model_name: str) -> bool:
-    # Bedrock rejects the cache marker for models LiteLLM's map doesn't
-    # recognise as cache-capable, so callers withhold it unless confirmed here.
-    import litellm
-
-    checker = getattr(getattr(litellm, "utils", None), "supports_prompt_caching", None)
-    for cand in _prompt_cache_name_candidates(model_name):
-        if checker is not None:
-            with contextlib.suppress(Exception):
-                if checker(cand):
-                    return True
-        entry = litellm.model_cost.get(cand)
-        if entry and entry.get("supports_prompt_caching"):
-            return True
-    return False

strix/config/settings.py

@@ -40,10 +40,6 @@ class LlmSettings(BaseSettings):
         default=False,
         alias="STRIX_FORCE_REQUIRED_TOOL_CHOICE",
     )
-    prompt_cache: bool = Field(
-        default=True,
-        alias="STRIX_PROMPT_CACHE",
-    )
     timeout: int = Field(default=300, alias="LLM_TIMEOUT")
 
 

strix/core/agents.py

@@ -14,15 +14,13 @@ from strix.core.sessions import session_write_lock
 
 
 if TYPE_CHECKING:
-    from collections.abc import Callable
-
     from agents.items import TResponseInputItem
     from agents.memory import Session
 
 
 logger = logging.getLogger(__name__)
 
-Status = Literal["running", "waiting", "completed", "stopped", "crashed", "failed", "budget_paused"]
+Status = Literal["running", "waiting", "completed", "stopped", "crashed", "failed"]
 
 
 @dataclass(slots=True)
@@ -49,9 +47,6 @@ class AgentCoordinator:
         self._snapshot_path: Path | None = None
         self.is_shutting_down = False
         self._budget_stopped = False
-        self._reserve_stopped = False
-        self._budget_paused = False
-        self._extend_budget: Callable[[], None] | None = None
 
     def set_snapshot_path(self, path: Path) -> None:
         self._snapshot_path = path
@@ -70,71 +65,6 @@ class AgentCoordinator:
             for runtime in self.runtimes.values():
                 runtime.wake.set()
 
-    @property
-    def reserve_stopped(self) -> bool:
-        return self._reserve_stopped
-
-    @property
-    def budget_paused(self) -> bool:
-        return self._budget_paused
-
-    def set_budget_extender(self, extend: Callable[[], None]) -> None:
-        self._extend_budget = extend
-
-    async def pause_for_budget(self, agent_id: str) -> None:
-        async with self._lock:
-            self._budget_paused = True
-        await self.set_status(agent_id, "budget_paused")
-
-    async def resume_from_budget_pause(self, *, exclude: str | None = None) -> None:
-        async with self._lock:
-            if not self._budget_paused:
-                return
-            self._budget_paused = False
-            paused = [aid for aid, status in self.statuses.items() if status == "budget_paused"]
-        if self._extend_budget is not None:
-            self._extend_budget()
-        for aid in paused:
-            await self.set_status(aid, "waiting")
-            if aid != exclude:
-                await self.send(
-                    aid,
-                    {
-                        "from": "system",
-                        "type": "budget_extended",
-                        "content": (
-                            "[Budget] The user extended the scan budget \u2014 continue your "
-                            "current task."
-                        ),
-                    },
-                )
-
-    async def reset_budget_stops(
-        self,
-        *,
-        budget_stopped: bool,
-        reserve_stopped: bool,
-        budget_paused: bool = False,
-    ) -> None:
-        async with self._lock:
-            self._budget_stopped = budget_stopped
-            self._reserve_stopped = reserve_stopped
-            if not budget_paused:
-                self._budget_paused = False
-                for aid, status in self.statuses.items():
-                    if status == "budget_paused":
-                        self.statuses[aid] = "waiting"
-        await self._maybe_snapshot()
-
-    async def claim_reserve_notification(self) -> str | None:
-        async with self._lock:
-            if self._reserve_stopped:
-                return None
-            self._reserve_stopped = True
-            for runtime in self.runtimes.values():
-                runtime.wake.set()
-            return next((aid for aid, parent in self.parent_of.items() if parent is None), None)
-
     async def register(
         self,
         agent_id: str,
@@ -202,8 +132,6 @@ class AgentCoordinator:
 
     async def send(self, target_agent_id: str, message: dict[str, Any]) -> bool:
         """Deliver a user/peer message by appending it to the target SDK session."""
-        if message.get("from") == "user" and self._budget_paused:
-            await self.resume_from_budget_pause(exclude=target_agent_id)
         async with self._lock:
             if target_agent_id not in self.statuses:
                 logger.debug("agent.send dropped unknown target=%s", target_agent_id)
@@ -238,8 +166,7 @@ class AgentCoordinator:
     async def wait_for_message(self, agent_id: str) -> None:
         while True:
             async with self._lock:
-                reserve_exit = self._reserve_stopped and self.parent_of.get(agent_id) is not None
-                if self._budget_stopped or reserve_exit or self.pending_counts.get(agent_id, 0) > 0:
+                if self._budget_stopped or self.pending_counts.get(agent_id, 0) > 0:
                     return
                 wake = self.runtimes.setdefault(agent_id, AgentRuntime()).wake
                 wake.clear()
@@ -373,9 +300,6 @@ class AgentCoordinator:
                 "metadata": {aid: dict(md) for aid, md in self.metadata.items()},
                 "pending_counts": dict(self.pending_counts),
                 "errors": dict(self.errors),
-                "budget_stopped": self._budget_stopped,
-                "reserve_stopped": self._reserve_stopped,
-                "budget_paused": self._budget_paused,
             }
 
     async def restore(self, snap: dict[str, Any]) -> None:
@@ -386,9 +310,6 @@ class AgentCoordinator:
             self.metadata = {aid: dict(md) for aid, md in snap.get("metadata", {}).items()}
             self.pending_counts = dict(snap.get("pending_counts", {}))
             self.errors = dict(snap.get("errors", {}))
-            self._budget_stopped = bool(snap.get("budget_stopped", False))
-            self._reserve_stopped = bool(snap.get("reserve_stopped", False))
-            self._budget_paused = bool(snap.get("budget_paused", False))
             for aid in self.statuses:
                 self.runtimes.setdefault(aid, AgentRuntime())
 

strix/core/execution.py

@@ -13,26 +13,15 @@ from agents import RunConfig, Runner
 from agents.exceptions import AgentsException, MaxTurnsExceeded, UserError
 from agents.sandbox.errors import ExecTransportError
 from docker import errors as docker_errors  # type: ignore[import-untyped, unused-ignore]
-from openai import (
-    APIConnectionError,
-    APIError,
-    APIStatusError,
-    APITimeoutError,
-    RateLimitError,
-)
+from openai import APIError
 
-from strix.core.hooks import (
-    BudgetExceededError,
-    BudgetPausedError,
-    SubagentBudgetReservedError,
-)
+from strix.core.hooks import BudgetExceededError
 from strix.core.inputs import child_initial_input
 from strix.core.sessions import (
     enforce_image_budget,
     open_agent_session,
     strip_all_images_from_session,
 )
-from strix.llm.compaction import is_context_overflow, maybe_compact
 
 
 if TYPE_CHECKING:
@@ -51,69 +40,6 @@ logger = logging.getLogger(__name__)
 StreamEventSink = Callable[[str, Any], None]
 
 _INPUT_REJECTION_CODES = frozenset({400, 404, 422})
-_MAX_COMPACTIONS_PER_CYCLE = 2
-
-
-def _run_config_model(run_config: RunConfig) -> str | None:
-    return run_config.model if isinstance(run_config.model, str) else None
-
-
-def _agent_instructions(agent: Any) -> str:
-    instructions = getattr(agent, "instructions", None)
-    return instructions if isinstance(instructions, str) else ""
-
-
-def _agent_tools_text(agent: Any) -> str:
-    parts: list[str] = []
-    for tool in getattr(agent, "tools", []) or []:
-        name = getattr(tool, "name", "")
-        description = getattr(tool, "description", "") or ""
-        schema = getattr(tool, "params_json_schema", "") or ""
-        parts.append(f"{name} {description} {schema}")
-    return "\n".join(parts)
-
-
-async def _compact_session(
-    agent: Any, session: Session, run_config: RunConfig, *, force: bool
-) -> bool:
-    model = _run_config_model(run_config)
-    if session is None or model is None:
-        return False
-    return await maybe_compact(
-        session,
-        model=model,
-        instructions=_agent_instructions(agent),
-        tools_text=_agent_tools_text(agent),
-        force=force,
-    )
-
-
-_TRANSIENT_MODEL_STATUS_CODES = frozenset({408, 500, 502, 503, 504})
-_MAX_TRANSIENT_MODEL_RETRIES = 4
-_TRANSIENT_MODEL_RETRY_BASE_DELAY_S = 2.0
-_TRANSIENT_MODEL_RETRY_MAX_DELAY_S = 30.0
-
-
-def _model_error_status_code(exc: BaseException) -> int | None:
-    code = getattr(exc, "status_code", None)
-    return code if isinstance(code, int) else None
-
-
-def _is_transient_model_error(exc: BaseException) -> bool:
-    if isinstance(exc, RateLimitError):
-        return False
-    if isinstance(exc, APITimeoutError | APIConnectionError):
-        return True
-    if isinstance(exc, APIStatusError):
-        return exc.status_code in _TRANSIENT_MODEL_STATUS_CODES
-    if isinstance(exc, APIError):
-        return _model_error_status_code(exc) is None
-    return False
-
-
-def _transient_model_retry_delay(attempt: int) -> float:
-    delay = _TRANSIENT_MODEL_RETRY_BASE_DELAY_S * float(2 ** (attempt - 1))
-    return min(delay, _TRANSIENT_MODEL_RETRY_MAX_DELAY_S)
 
 
 async def run_agent_loop(
@@ -138,34 +64,21 @@ async def run_agent_loop(
     )
     result: RunResultBase | None = None
 
-    budget_stopped = coordinator.budget_stopped
-    reserve_stopped = coordinator.reserve_stopped
-    if budget_stopped:
-        await coordinator.set_status(agent_id, "stopped")
-        raise BudgetExceededError("scan budget reached")
-    if reserve_stopped and context.get("parent_id") is not None:
-        await coordinator.set_status(agent_id, "stopped")
-        raise SubagentBudgetReservedError("scan reached the sub-agent budget reserve")
-
-    if reserve_stopped and start_parked and interactive and context.get("parent_id") is None:
-        await coordinator.send(agent_id, _reserve_notice())
-
     if not (start_parked and interactive):
         if interactive:
-            with contextlib.suppress(BudgetPausedError):
-                result = await _run_cycle(
-                    agent,
-                    coordinator,
-                    agent_id,
-                    input_data=initial_input,
-                    run_config=run_config,
-                    context=context,
-                    max_turns=max_turns,
-                    session=session,
-                    interactive=interactive,
-                    event_sink=event_sink,
-                    hooks=hooks,
-                )
+            result = await _run_cycle(
+                agent,
+                coordinator,
+                agent_id,
+                input_data=initial_input,
+                run_config=run_config,
+                context=context,
+                max_turns=max_turns,
+                session=session,
+                interactive=interactive,
+                event_sink=event_sink,
+                hooks=hooks,
+            )
         else:
             result = await _run_noninteractive_until_lifecycle(
                 agent,
@@ -193,25 +106,20 @@ async def run_agent_loop(
             await coordinator.set_status(agent_id, "stopped")
             raise BudgetExceededError("scan budget reached")
 
-        if coordinator.reserve_stopped and context.get("parent_id") is not None:
-            await coordinator.set_status(agent_id, "stopped")
-            raise SubagentBudgetReservedError("scan reached the sub-agent budget reserve")
-
         await coordinator.consume_pending(agent_id)
-        with contextlib.suppress(BudgetPausedError):
-            result = await _run_cycle(
-                agent,
-                coordinator,
-                agent_id,
-                input_data=[],
-                run_config=run_config,
-                context=context,
-                max_turns=max_turns,
-                session=session,
-                interactive=interactive,
-                event_sink=event_sink,
-                hooks=hooks,
-            )
+        result = await _run_cycle(
+            agent,
+            coordinator,
+            agent_id,
+            input_data=[],
+            run_config=run_config,
+            context=context,
+            max_turns=max_turns,
+            session=session,
+            interactive=interactive,
+            event_sink=event_sink,
+            hooks=hooks,
+        )
 
 
 async def spawn_child_agent(
@@ -383,10 +291,6 @@ async def _run_noninteractive_until_lifecycle(
             await coordinator.set_status(agent_id, "stopped")
             raise BudgetExceededError("scan budget reached")
 
-        if coordinator.reserve_stopped and context.get("parent_id") is not None:
-            await coordinator.set_status(agent_id, "stopped")
-            raise SubagentBudgetReservedError("scan reached the sub-agent budget reserve")
-
         result = await _run_cycle(
             agent,
             coordinator,
@@ -417,7 +321,7 @@ async def _run_noninteractive_until_lifecycle(
 
         if invalid_final_outputs >= invalid_final_output_limit:
             await coordinator.set_status(agent_id, "crashed")
-            await _notify_parent_on_terminal(coordinator, agent_id, "crashed")
+            await _notify_parent_on_crash(coordinator, agent_id, "crashed")
             raise MaxTurnsExceeded(
                 "Agent exhausted non-interactive recovery attempts without calling "
                 "finish_scan or agent_finish."
@@ -446,8 +350,6 @@ async def _run_cycle(  # noqa: PLR0912, PLR0915
     hooks: RunHooks[dict[str, Any]] | None,
 ) -> RunResultBase | None:
     image_strips = 0
-    compactions = 0
-    model_retries = 0
     while True:
         try:
             await coordinator.mark_running(agent_id)
@@ -458,10 +360,6 @@ async def _run_cycle(  # noqa: PLR0912, PLR0915
                         await enforce_image_budget(session, max_images)
                     except Exception:
                         logger.exception("image-budget enforcement failed for %s", agent_id)
-                try:
-                    await _compact_session(agent, session, run_config, force=False)
-                except Exception:
-                    logger.exception("proactive compaction failed for %s", agent_id)
             stream = Runner.run_streamed(
                 agent,
                 input=input_data,
@@ -482,7 +380,9 @@ async def _run_cycle(  # noqa: PLR0912, PLR0915
                                 logger.exception("stream event sink failed for %s", agent_id)
                     if stream.run_loop_exception is not None:
                         raise stream.run_loop_exception
-                except (BudgetExceededError, BudgetPausedError, SubagentBudgetReservedError):
+                except BudgetExceededError:
+                    # A RuntimeError subclass: re-raise explicitly so it is never
+                    # mistaken for the LiteLLM "after shutdown" race below.
                     raise
                 except RuntimeError as stream_exc:
                     if "after shutdown" not in str(stream_exc):
@@ -501,15 +401,6 @@ async def _run_cycle(  # noqa: PLR0912, PLR0915
                     )
             finally:
                 await coordinator.detach_stream(agent_id, stream)
-        except BudgetPausedError as exc:
-            logger.info("agent %s paused at the scan budget limit: %s", agent_id, exc)
-            await coordinator.pause_for_budget(agent_id)
-            raise
-        except SubagentBudgetReservedError as exc:
-            logger.info("sub-agent %s stopped at the budget reserve: %s", agent_id, exc)
-            await coordinator.set_status(agent_id, "stopped")
-            await _notify_root_on_budget_reserve(coordinator)
-            raise
         except BudgetExceededError as exc:
             logger.info(
                 "agent %s reached the scan budget limit; stopping the scan: %s", agent_id, exc
@@ -537,41 +428,6 @@ async def _run_cycle(  # noqa: PLR0912, PLR0915
                     )
                     input_data = []
                     continue
-            if (
-                compactions < _MAX_COMPACTIONS_PER_CYCLE
-                and session is not None
-                and is_context_overflow(exc)
-            ):
-                try:
-                    compacted = await _compact_session(agent, session, run_config, force=True)
-                except Exception:
-                    logger.exception("overflow compaction recovery failed for %s", agent_id)
-                    compacted = False
-                if compacted:
-                    compactions += 1
-                    logger.info(
-                        "Compacted %s session after context overflow; retrying (%d)",
-                        agent_id,
-                        compactions,
-                    )
-                    input_data = []
-                    continue
-            if model_retries < _MAX_TRANSIENT_MODEL_RETRIES and _is_transient_model_error(exc):
-                model_retries += 1
-                delay = _transient_model_retry_delay(model_retries)
-                logger.warning(
-                    "transient model/provider error for %s; replaying turn "
-                    "(attempt %d/%d, backoff %.1fs): %r",
-                    agent_id,
-                    model_retries,
-                    _MAX_TRANSIENT_MODEL_RETRIES,
-                    delay,
-                    exc,
-                )
-                await asyncio.sleep(delay)
-                if session is not None:
-                    input_data = []
-                continue
             if not interactive:
                 raise
             if isinstance(exc, MaxTurnsExceeded):
@@ -582,7 +438,7 @@ async def _run_cycle(  # noqa: PLR0912, PLR0915
                 status = "crashed"
             logger.exception("agent run failed for %s; parking as %s", agent_id, status)
             await coordinator.set_status(agent_id, status, error=str(exc) or type(exc).__name__)
-            await _notify_parent_on_terminal(coordinator, agent_id, status)
+            await _notify_parent_on_crash(coordinator, agent_id, status)
             return None
         else:
             await _settle_run_result(coordinator, agent_id, interactive)
@@ -646,31 +502,12 @@ async def _append_noninteractive_tool_required_message(
     return []
 
 
-_TERMINAL_NOTICE = {
-    "crashed": (
-        "[Agent crash] {name} ({agent_id}) terminated unexpectedly. "
-        "Stop waiting on this child unless you want to message it again."
-    ),
-    "failed": (
-        "[Agent failed] {name} ({agent_id}) stopped with an error and will not "
-        "send a completion report. Stop waiting on this child unless you want to "
-        "message it again."
-    ),
-    "stopped": (
-        "[Agent capped] {name} ({agent_id}) hit its turn limit and was stopped "
-        "before finishing. It will not send a completion report, so stop waiting "
-        "on this child; account for its capped subtask and continue."
-    ),
-}
-
-
-async def _notify_parent_on_terminal(
+async def _notify_parent_on_crash(
     coordinator: AgentCoordinator,
     agent_id: str,
     status: str,
 ) -> None:
-    template = _TERMINAL_NOTICE.get(status)
-    if template is None:
+    if status != "crashed":
         return
     async with coordinator._lock:
         parent = coordinator.parent_of.get(agent_id)
@@ -681,35 +518,16 @@ async def _notify_parent_on_terminal(
         parent,
         {
             "from": agent_id,
-            "type": status,
+            "type": "crash",
             "priority": "high",
-            "content": template.format(name=name, agent_id=agent_id),
+            "content": (
+                f"[Agent crash] {name} ({agent_id}) terminated unexpectedly. "
+                "Stop waiting on this child unless you want to message it again."
+            ),
         },
     )
 
 
-def _reserve_notice() -> dict[str, Any]:
-    return {
-        "from": "system",
-        "type": "budget_reserve_stop",
-        "priority": "high",
-        "content": (
-            "[Budget reserve] The scan has reached the sub-agent budget reserve: every "
-            "sub-agent is being force-stopped as soon as its in-flight turn completes, and "
-            "none will send a completion report. Their confirmed vulnerabilities are "
-            "already filed as they were found. Do not wait on any sub-agents and do not "
-            "spawn new ones — wrap up now and call finish_scan."
-        ),
-    }
-
-
-async def _notify_root_on_budget_reserve(coordinator: AgentCoordinator) -> None:
-    root = await coordinator.claim_reserve_notification()
-    if root is None:
-        return
-    await coordinator.send(root, _reserve_notice())
-
-
 async def _start_child_runner(
     *,
     parent_ctx: dict[str, Any],
@@ -761,8 +579,6 @@ async def _start_child_runner(
             )
         except BudgetExceededError:
             logger.info("child %s stopped after reaching the scan budget limit", child_id)
-        except SubagentBudgetReservedError:
-            logger.info("child %s stopped at the sub-agent budget reserve", child_id)
 
     task_handle = asyncio.create_task(_child_loop(), name=f"agent-{name}-{child_id}")
     await coordinator.attach_runtime(child_id, task=task_handle)

strix/core/hooks.py

@@ -14,210 +14,26 @@ from strix.report.state import get_global_report_state
 if TYPE_CHECKING:
     from agents import RunContextWrapper
     from agents.agent import Agent
-    from agents.items import ModelResponse, TResponseInputItem
+    from agents.items import ModelResponse
 
 
 logger = logging.getLogger(__name__)
 
 
-_STAGE_LABELS: tuple[str, ...] = ("NOTICE", "URGENT", "CRITICAL")
-_TURN_WARN_BANDS: tuple[float, ...] = (0.70, 0.85, 0.95)
-_ROOT_BUDGET_WARN_BANDS: tuple[float, ...] = (0.70, 0.85, 0.95)
-_SUBAGENT_BUDGET_WARN_BANDS: tuple[float, ...] = (0.75, 0.80, 0.85)
-_SUBAGENT_BUDGET_RESERVE = 0.90
-
-
 class BudgetExceededError(RuntimeError):
     """Raised when the accumulated LLM cost reaches the configured budget."""
 
 
-class SubagentBudgetReservedError(RuntimeError):
-    """Raised to stop a single sub-agent once the reserve threshold is crossed."""
-
-
-class BudgetPausedError(RuntimeError):
-    """Raised to park one agent when an interactive scan reaches its budget."""
-
-
-def recomputed_budget_flags(
-    cost: float,
-    max_budget_usd: float | None,
-    *,
-    interactive: bool,
-) -> tuple[bool, bool]:
-    """Return the (budget_stopped, reserve_stopped) flags a resumed scan should carry."""
-    if max_budget_usd is None:
-        return False, False
-    if interactive:
-        return False, False
-    budget_stopped = cost >= max_budget_usd
-    reserve_stopped = cost >= max_budget_usd * _SUBAGENT_BUDGET_RESERVE
-    return budget_stopped, reserve_stopped
-
-
-def _crossed_stage(fraction: float, bands: tuple[float, ...]) -> int | None:
-    crossed: int | None = None
-    for index, band in enumerate(bands):
-        if fraction >= band:
-            crossed = index
-    return crossed
-
-
-_ROOT_DIRECTIVES: tuple[str, ...] = (
-    (
-        "As the root agent, begin planning your wind-down of the whole scan: avoid "
-        "starting large new lines of investigation, and keep your required objectives on "
-        "track so you can call finish_scan comfortably before the limit."
-    ),
-    (
-        "As the root agent, prioritize wrapping up the whole scan now: stop opening new "
-        "lines of investigation, close out only what is essential, and move toward calling "
-        "finish_scan to compile and deliver the final report."
-    ),
-    (
-        "As the root agent, STOP all other work on the whole scan and finish immediately: "
-        "secure your findings and call finish_scan now — anything left unfinished when the "
-        "limit is hit is discarded."
-    ),
-)
-_SUBAGENT_DIRECTIVES: tuple[str, ...] = (
-    (
-        "As a sub-agent, begin planning your wind-down: avoid starting large new subtasks, "
-        "and if you are close to a confirmed, validated vulnerability, drive it to a result "
-        "you can report."
-    ),
-    (
-        "As a sub-agent, prioritize wrapping up your task now: report any confirmed, "
-        "validated vulnerability, finish work that is nearly done rather than starting "
-        "anything new, and prepare to call agent_finish."
-    ),
-    (
-        "As a sub-agent, STOP all other work and finish immediately: report any confirmed "
-        "vulnerability right now and call agent_finish to hand your results back to your "
-        "parent before you are cut off."
-    ),
-)
-
-
-def _wrapup_directive(context: RunContextWrapper[dict[str, Any]], stage: int) -> str:
-    is_root = context.context.get("parent_id") is None
-    directives = _ROOT_DIRECTIVES if is_root else _SUBAGENT_DIRECTIVES
-    return directives[stage]
-
-
-def _urgency(stage: int) -> str:
-    return _STAGE_LABELS[stage]
-
-
 class ReportUsageHooks(RunHooks[dict[str, Any]]):
-    """Persist SDK-native usage and warn/stop as turn and cost budgets are consumed."""
+    """Persist SDK-native usage after every model response."""
 
-    def __init__(
-        self,
-        *,
-        model: str,
-        max_budget_usd: float | None = None,
-        max_turns: int | None = None,
-        interactive: bool = False,
-    ) -> None:
+    def __init__(self, *, model: str, max_budget_usd: float | None = None) -> None:
         if max_budget_usd is not None and (
             not math.isfinite(max_budget_usd) or max_budget_usd <= 0
         ):
             raise ValueError("max_budget_usd must be a finite number greater than 0")
-        if max_turns is not None and max_turns <= 0:
-            raise ValueError("max_turns must be a positive integer")
         self._model = model
         self._max_budget_usd = max_budget_usd
-        self._budget_increment = max_budget_usd
-        self._max_turns = max_turns
-        self._interactive = interactive
-
-    def extend_budget(self) -> None:
-        if self._max_budget_usd is None or self._budget_increment is None:
-            return
-        self._max_budget_usd += self._budget_increment
-
-    async def on_llm_start(
-        self,
-        context: RunContextWrapper[dict[str, Any]],
-        agent: Agent[dict[str, Any]],  # noqa: ARG002
-        system_prompt: str | None,  # noqa: ARG002
-        input_items: list[TResponseInputItem],
-    ) -> None:
-        try:
-            self._maybe_warn_turns(context, input_items)
-            self._maybe_warn_budget(context, input_items)
-        except Exception:
-            logger.exception("budget/turn warning injection failed")
-
-    def _maybe_warn_turns(
-        self,
-        context: RunContextWrapper[dict[str, Any]],
-        input_items: list[TResponseInputItem],
-    ) -> None:
-        if not self._max_turns:
-            return
-        usage = getattr(context, "usage", None)
-        requests = getattr(usage, "requests", None)
-        if not isinstance(requests, int):
-            return
-        turns_used = requests + 1
-        stage = _crossed_stage(turns_used / self._max_turns, _TURN_WARN_BANDS)
-        if stage is None:
-            return
-        remaining = max(self._max_turns - turns_used, 0)
-        pct = round(100 * turns_used / self._max_turns)
-        content = (
-            f"[{_urgency(stage)}] Turn budget: {turns_used}/{self._max_turns} used ({pct}%). "
-            f"About {remaining} turn(s) remain before this agent is force-stopped and any "
-            f"in-progress work is discarded. {_wrapup_directive(context, stage)}"
-        )
-        input_items.append({"role": "user", "content": content})
-
-    def _maybe_warn_budget(
-        self,
-        context: RunContextWrapper[dict[str, Any]],
-        input_items: list[TResponseInputItem],
-    ) -> None:
-        if self._max_budget_usd is None:
-            return
-        report_state = get_global_report_state()
-        if report_state is None:
-            return
-        cost = report_state.get_total_llm_cost()
-        is_root = context.context.get("parent_id") is None
-        if self._interactive:
-            bands = _ROOT_BUDGET_WARN_BANDS
-        else:
-            bands = _ROOT_BUDGET_WARN_BANDS if is_root else _SUBAGENT_BUDGET_WARN_BANDS
-        stage = _crossed_stage(cost / self._max_budget_usd, bands)
-        if stage is None:
-            return
-        pct = round(100 * cost / self._max_budget_usd)
-        reserve_pct = round(_SUBAGENT_BUDGET_RESERVE * 100)
-        if self._interactive:
-            content = (
-                f"[{_urgency(stage)}] Scan cost budget: ${cost:.2f}/${self._max_budget_usd:.2f} "
-                f"spent ({pct}%). This budget is shared across every agent in the scan; when it "
-                "is reached all agents are paused until the user chooses to continue. "
-                f"{_wrapup_directive(context, stage)}"
-            )
-        elif is_root:
-            content = (
-                f"[{_urgency(stage)}] Scan cost budget: ${cost:.2f}/${self._max_budget_usd:.2f} "
-                f"spent ({pct}%). This budget is shared across every agent in the scan; when it "
-                "is reached the whole scan is stopped immediately, and sub-agents are stopped at "
-                f"{reserve_pct}% to reserve the remainder for your final report. "
-                f"{_wrapup_directive(context, stage)}"
-            )
-        else:
-            content = (
-                f"[{_urgency(stage)}] Scan cost budget: ${cost:.2f}/${self._max_budget_usd:.2f} "
-                f"spent ({pct}%). This budget is shared across every agent in the scan; "
-                f"sub-agents are stopped at {reserve_pct}% to leave the remainder for the root "
-                f"agent's final report. {_wrapup_directive(context, stage)}"
-            )
-        input_items.append({"role": "user", "content": content})
 
     async def on_llm_end(
         self,
@@ -250,21 +66,6 @@ class ReportUsageHooks(RunHooks[dict[str, Any]]):
         if self._max_budget_usd is not None:
             cost = report_state.get_total_llm_cost()
             if cost >= self._max_budget_usd:
-                if self._interactive:
-                    raise BudgetPausedError(
-                        f"Scan budget of ${self._max_budget_usd:.2f} reached "
-                        f"(spent ${cost:.4f}); pausing until the user continues"
-                    )
                 raise BudgetExceededError(
                     f"Token budget of ${self._max_budget_usd:.2f} exceeded (spent ${cost:.4f})"
                 )
-            is_root = ctx.get("parent_id") is None
-            if not self._interactive and not is_root:
-                reserve_limit = self._max_budget_usd * _SUBAGENT_BUDGET_RESERVE
-                if cost >= reserve_limit:
-                    raise SubagentBudgetReservedError(
-                        f"Sub-agent budget reserve reached: spent ${cost:.4f} of "
-                        f"${self._max_budget_usd:.2f} "
-                        f"(>= {round(_SUBAGENT_BUDGET_RESERVE * 100)}% reserve); stopping this "
-                        "sub-agent so the root agent can finish the scan."
-                    )

strix/core/inputs.py

@@ -10,9 +10,6 @@ from openai.types.shared import Reasoning
 
 from strix.config.models import (
     DEFAULT_MODEL_RETRY,
-    bedrock_route_supports_prompt_caching,
-    is_bedrock_route,
-    is_claude_model,
     is_known_openai_bare_model,
     model_supports_reasoning,
     request_timeout_extra_args,
@@ -131,7 +128,6 @@ def make_model_settings(
     model_name: str,
     force_required_tool_choice: bool = False,
     request_timeout: float | None = None,
-    prompt_cache: bool = True,
 ) -> ModelSettings:
     model_settings = ModelSettings(
         parallel_tool_calls=False,
@@ -149,38 +145,9 @@ def make_model_settings(
         )
     if force_required_tool_choice and _accepts_required_tool_choice(model_name):
         model_settings = model_settings.resolve(ModelSettings(tool_choice="required"))
-
-    cache_extra_args = _prompt_cache_extra_args(model_name) if prompt_cache else None
-    if cache_extra_args:
-        model_settings = model_settings.resolve(
-            ModelSettings(
-                extra_args={**(model_settings.extra_args or {}), **cache_extra_args},
-            ),
-        )
     return model_settings
 
 
-def _prompt_cache_extra_args(model_name: str) -> dict[str, Any] | None:
-    """LiteLLM ``cache_control_injection_points`` for Claude prompt caching.
-
-    System prompt + rolling last-message breakpoint everywhere; ``tool_config``
-    only on Bedrock Converse (the only route whose LiteLLM transform consumes
-    it — elsewhere it leaks onto the wire and native Anthropic 400s). Unmapped
-    Bedrock models get no points at all: Bedrock rejects the passed-through
-    field outright.
-    """
-    if not is_claude_model(model_name):
-        return None
-    if is_bedrock_route(model_name) and not bedrock_route_supports_prompt_caching(model_name):
-        return None
-
-    points: list[dict[str, Any]] = [{"location": "message", "role": "system"}]
-    if is_bedrock_route(model_name):
-        points.append({"location": "tool_config"})
-    points.append({"location": "message", "index": -1})
-    return {"cache_control_injection_points": points}
-
-
 def child_initial_input(
     *,
     name: str,

strix/core/runner.py

@@ -3,12 +3,10 @@
 from __future__ import annotations
 
 import contextlib
-import io
 import json
 import logging
 import uuid
 from collections.abc import Callable
-from pathlib import Path
 from typing import TYPE_CHECKING, Any
 
 from agents import RunConfig
@@ -31,7 +29,7 @@ from strix.core.execution import (
 from strix.core.execution import (
     spawn_child_agent as start_child_agent,
 )
-from strix.core.hooks import BudgetExceededError, ReportUsageHooks, recomputed_budget_flags
+from strix.core.hooks import BudgetExceededError, ReportUsageHooks
 from strix.core.inputs import (
     DEFAULT_MAX_TURNS,
     build_root_task,
@@ -40,13 +38,8 @@ from strix.core.inputs import (
 )
 from strix.core.paths import run_dir_for, runtime_state_dir
 from strix.core.sessions import open_agent_session
-from strix.report.state import get_global_report_state
 from strix.runtime import session_manager
 from strix.telemetry.logging import set_scan_id, setup_scan_logging
-from strix.tools.output_store import (
-    WORKSPACE_SPILL_DIR,
-    configure_spill_writer,
-)
 
 
 if TYPE_CHECKING:
@@ -186,18 +179,6 @@ async def run_strix_scan(
                 f"Cannot resume scan {scan_id}: missing SDK session database at {agents_db}",
             )
         await coordinator.restore(snap)
-        report_state = get_global_report_state()
-        if report_state is not None:
-            budget_stopped, reserve_stopped = recomputed_budget_flags(
-                report_state.get_total_llm_cost(),
-                max_budget_usd,
-                interactive=interactive,
-            )
-            await coordinator.reset_budget_stops(
-                budget_stopped=budget_stopped,
-                reserve_stopped=reserve_stopped,
-                budget_paused=interactive and coordinator.budget_paused,
-            )
         for aid, parent in coordinator.parent_of.items():
             if parent is None:
                 root_id = aid
@@ -222,20 +203,6 @@ async def run_strix_scan(
     )
     logger.info("Sandbox ready for scan %s", scan_id)
 
-    sandbox_session = bundle["session"]
-
-    async def _spill_to_workspace(output_id: str, text: str) -> str | None:
-        """Write an oversized tool result into the sandbox; return its path or None."""
-        path = f"{WORKSPACE_SPILL_DIR}/{output_id}.txt"
-        try:
-            await sandbox_session.write(Path(path), io.BytesIO(text.encode("utf-8")))
-        except Exception:
-            logger.exception("failed to spill tool output to sandbox workspace")
-            return None
-        return path
-
-    configure_spill_writer(_spill_to_workspace)
-
     sessions_to_close: list[SQLiteSession] = []
 
     try:
@@ -249,7 +216,6 @@ async def run_strix_scan(
             model_name=resolved_model,
             force_required_tool_choice=settings.llm.force_required_tool_choice,
             request_timeout=settings.llm.timeout,
-            prompt_cache=settings.llm.prompt_cache,
         )
         run_config = RunConfig(
             model=resolved_model,
@@ -258,14 +224,7 @@ async def run_strix_scan(
             sandbox=SandboxRunConfig(client=bundle["client"], session=bundle["session"]),
             trace_include_sensitive_data=False,
         )
-        hooks = ReportUsageHooks(
-            model=resolved_model,
-            max_budget_usd=max_budget_usd,
-            max_turns=max_turns,
-            interactive=interactive,
-        )
-        if interactive:
-            coordinator.set_budget_extender(hooks.extend_budget)
+        hooks = ReportUsageHooks(model=resolved_model, max_budget_usd=max_budget_usd)
 
         scope_context = build_scope_context(scan_config)
         root_context = _merge_root_prompt_context(scope_context, extra_system_prompt_context)
@@ -440,7 +399,6 @@ async def run_strix_scan(
                 await coordinator.set_status(root_id, "failed")
         raise
     finally:
-        configure_spill_writer(None)
         for s in sessions_to_close:
             with contextlib.suppress(Exception):
                 s.close()

strix/core/sessions.py

@@ -92,39 +92,6 @@ async def _rewrite_session(
         return True
 
 
-async def replace_session_items(
-    session: Session,
-    new_items: list[Any],
-    *,
-    expected_len: int | None = None,
-) -> bool:
-    """Overwrite the session's items, restoring the originals on failure.
-
-    When ``expected_len`` is given, the rewrite is skipped if the session no
-    longer has that many items (a concurrent writer changed it), so a slow
-    compaction summary can't clobber newer turns.
-    """
-    async with session_write_lock(session):
-        original = list(await session.get_items())
-        if expected_len is not None and len(original) != expected_len:
-            logger.warning(
-                "skipping session rewrite: expected %d items, found %d",
-                expected_len,
-                len(original),
-            )
-            return False
-        rebuilt = cast("list[TResponseInputItem]", new_items)
-        await session.clear_session()
-        try:
-            await session.add_items(rebuilt)
-        except Exception:
-            logger.exception("session rewrite failed; restoring original items")
-            await session.clear_session()
-            await session.add_items(original)
-            raise
-        return True
-
-
 async def strip_all_images_from_session(session: Session) -> bool:
     """Replace every image tool output with a text placeholder (rejection recovery)."""
 

strix/interface/cli.py

@@ -13,7 +13,6 @@ from rich.panel import Panel
 from rich.text import Text
 
 from strix.config import load_settings
-from strix.core.inputs import DEFAULT_MAX_TURNS
 from strix.core.runner import run_strix_scan
 from strix.report.state import ReportState, set_global_report_state
 from strix.runtime import session_manager
@@ -185,7 +184,6 @@ async def run_cli(args: Any) -> None:  # noqa: PLR0915
                     local_sources=getattr(args, "local_sources", None) or [],
                     interactive=bool(getattr(args, "interactive", False)),
                     max_budget_usd=getattr(args, "max_budget_usd", None),
-                    max_turns=getattr(args, "max_turns", DEFAULT_MAX_TURNS),
                 )
             finally:
                 stop_updates.set()

strix/interface/main.py

@@ -31,7 +31,6 @@ from strix.config.models import (
     is_known_openai_bare_model,
     is_recommended_or_frontier_model,
 )
-from strix.core.inputs import DEFAULT_MAX_TURNS
 from strix.core.paths import run_dir_for, runtime_state_dir
 from strix.interface.cli import run_cli
 from strix.interface.tui import run_tui
@@ -211,7 +210,7 @@ def validate_environment() -> None:
             padding=(1, 2),
         )
 
-        logger.debug("Missing required env vars: %s", missing_required_vars)
+        logger.error("Missing required env vars: %s", missing_required_vars)
         console.print("\n")
         console.print(panel)
         console.print()
@@ -224,7 +223,7 @@ def validate_environment() -> None:
 
 def check_docker_installed() -> None:
     if shutil.which("docker") is None:
-        logger.debug("Docker CLI not found in PATH")
+        logger.error("Docker CLI not found in PATH")
         console = Console()
         error_text = Text()
         error_text.append("DOCKER NOT INSTALLED", style="bold red")
@@ -423,7 +422,7 @@ async def warm_up_llm(show_model_warning: bool = True) -> None:
             logger.info("LLM warm-up succeeded for dedupe model %s", dedupe_model)
 
     except Exception as e:
-        logger.debug("LLM warm-up failed", exc_info=True)
+        logger.exception("LLM warm-up failed")
         error_text = Text()
         sub_hint = _subscription_error_hint(e)
         if sub_hint is not None:
@@ -482,16 +481,6 @@ def _positive_budget(value: str) -> float:
     return budget
 
 
-def _positive_int(value: str) -> int:
-    try:
-        parsed = int(value)
-    except ValueError as exc:
-        raise argparse.ArgumentTypeError(f"invalid int value: {value!r}") from exc
-    if parsed <= 0:
-        raise argparse.ArgumentTypeError("must be an integer greater than 0")
-    return parsed
-
-
 def parse_arguments() -> argparse.Namespace:
     parser = argparse.ArgumentParser(
         description="Strix Multi-Agent Cybersecurity Penetration Testing Tool",
@@ -647,27 +636,10 @@ Examples:
     )
 
     parser.add_argument(
-        "--max-budget",
-        dest="max_budget_usd",
-        metavar="USD",
+        "--max-budget-usd",
         type=_positive_budget,
         default=None,
-        help=(
-            "Maximum LLM cost in USD (> 0). The scan stops cleanly when this limit is reached. "
-            "Graduated wrap-up warnings are sent to all agents as it is approached."
-        ),
-    )
-
-    parser.add_argument(
-        "--max-turns",
-        dest="max_turns",
-        metavar="N",
-        type=_positive_int,
-        default=DEFAULT_MAX_TURNS,
-        help=(
-            "Maximum turns per agent (> 0, default %(default)s). Each agent is force-stopped "
-            "when it reaches this limit, with graduated wrap-up warnings as it is approached."
-        ),
+        help="Maximum LLM cost in USD (> 0). The scan stops cleanly when this limit is reached.",
     )
 
     parser.add_argument(
@@ -946,7 +918,7 @@ def pull_docker_image() -> None:
                 last_update = process_pull_line(line, layers_info, status, last_update)
 
         except DockerException as e:
-            logger.debug("Failed to pull docker image %s", image, exc_info=True)
+            logger.exception("Failed to pull docker image %s", image)
             console.print()
             error_text = Text()
             error_text.append("FAILED TO PULL IMAGE", style="bold red")
@@ -980,7 +952,7 @@ def main() -> None:
     # `strix view [<run>]` is a viewer-only subcommand, dispatched before the
     # scan argument parser (which requires a target) and before any scan setup.
     if len(sys.argv) > 1 and sys.argv[1] == "view":
-        from strix.interface.viewer.cli import run_view
+        from strix.viewer.cli import run_view
 
         run_view(sys.argv[2:])
         return

strix/interface/tui/app.py

@@ -34,7 +34,6 @@ from textual.widgets.tree import TreeNode
 from strix.config import load_settings
 from strix.config.models import is_recommended_or_frontier_model
 from strix.core.hooks import BudgetExceededError
-from strix.core.inputs import DEFAULT_MAX_TURNS
 from strix.core.runner import run_strix_scan
 from strix.interface.tui.live_view import TuiLiveView
 from strix.interface.tui.messages import send_user_message_to_agent
@@ -815,7 +814,6 @@ class StrixTUIApp(App):  # type: ignore[misc]
         self._scan_completed = threading.Event()
         self._scan_error: BaseException | None = None
         self._error_noted_agents: set[str] = set()
-        self._budget_pause_notified = False
 
         self._spinner_frame_index: int = 0
         self._sweep_num_squares: int = 6
@@ -1048,7 +1046,6 @@ class StrixTUIApp(App):  # type: ignore[misc]
                                 self.live_view.record_agent_error(agent_id, error)
                         else:
                             self._error_noted_agents.discard(agent_id)
-                    self._notify_budget_pause(statuses)
 
         if self._scan_loop is None or self._scan_loop.is_closed():
             return
@@ -1060,19 +1057,6 @@ class StrixTUIApp(App):  # type: ignore[misc]
 
         self._agent_graph_sync_future = asyncio.run_coroutine_threadsafe(collect(), self._scan_loop)
 
-    def _notify_budget_pause(self, statuses: dict[str, Any]) -> None:
-        paused = any(status == "budget_paused" for status in statuses.values())
-        if paused and not self._budget_pause_notified:
-            self._budget_pause_notified = True
-            self.notify(
-                "Budget limit reached \u2014 agents paused. Send a message to continue "
-                "(this extends the budget), or ctrl-q to quit.",
-                severity="warning",
-                timeout=15,
-            )
-        elif not paused:
-            self._budget_pause_notified = False
-
     def _update_agent_node(self, agent_id: str, agent_data: dict[str, Any]) -> bool:
         if agent_id not in self.agent_nodes:
             return False
@@ -1085,7 +1069,6 @@ class StrixTUIApp(App):  # type: ignore[misc]
             status_indicators = {
                 "running": "⚪",
                 "waiting": "⏸",
-                "budget_paused": "⏸",
                 "completed": "🟢",
                 "failed": "🔴",
                 "crashed": "🔴",
@@ -1283,17 +1266,10 @@ class StrixTUIApp(App):  # type: ignore[misc]
             self._stop_dot_animation()
             return (text, Text(), False)
 
-        if status in {"waiting", "budget_paused"}:
+        if status == "waiting":
             text = Text()
-            keymap = Text()
-            if status == "budget_paused":
-                text.append("Budget limit reached", style="yellow")
-                text.append(" \u00b7 ", style="dim")
-                text.append("Send a message to continue", style="dim")
-                keymap = keymap_styled([("ctrl-q", "quit")])
-            else:
-                text.append("Send message to resume", style="dim")
-            return (text, keymap, False)
+            text.append("Send message to resume", style="dim")
+            return (text, Text(), False)
 
         if status == "running":
             if self._agent_has_real_activity(agent_id):
@@ -1518,7 +1494,6 @@ class StrixTUIApp(App):  # type: ignore[misc]
                                 coordinator=self.coordinator,
                                 interactive=True,
                                 max_budget_usd=getattr(self.args, "max_budget_usd", None),
-                                max_turns=getattr(self.args, "max_turns", DEFAULT_MAX_TURNS),
                                 event_sink=self._capture_sdk_event,
                             ),
                         )
@@ -1526,7 +1501,10 @@ class StrixTUIApp(App):  # type: ignore[misc]
                 except (KeyboardInterrupt, asyncio.CancelledError):
                     logger.info("Scan interrupted by user")
                 except BudgetExceededError:
-                    logger.info("Scan stopped: --max-budget limit reached")
+                    # Defensive: the runner stops the scan cleanly on budget and
+                    # returns, so this normally never propagates. Treat it as a
+                    # graceful stop, not a scan error, if it ever does.
+                    logger.info("Scan stopped: --max-budget-usd limit reached")
                 except (ConnectionError, TimeoutError) as e:
                     logging.exception("Network error during scan")
                     self._scan_error = e
@@ -1581,7 +1559,6 @@ class StrixTUIApp(App):  # type: ignore[misc]
         status_indicators = {
             "running": "⚪",
             "waiting": "⏸",
-            "budget_paused": "⏸",
             "completed": "🟢",
             "failed": "🔴",
             "crashed": "🔴",
@@ -1628,7 +1605,6 @@ class StrixTUIApp(App):  # type: ignore[misc]
         status_indicators = {
             "running": "⚪",
             "waiting": "⏸",
-            "budget_paused": "⏸",
             "completed": "🟢",
             "failed": "🔴",
             "crashed": "🔴",
@@ -1753,10 +1729,7 @@ class StrixTUIApp(App):  # type: ignore[misc]
             message=message,
         )
         if not submitted:
-            if self._scan_completed.is_set():
-                self.notify("The scan has ended; message was not sent", severity="warning")
-            else:
-                self.notify("Scan loop is not ready; message was not sent", severity="warning")
+            self.notify("Scan loop is not ready; message was not sent", severity="warning")
             return
 
         self._displayed_events.clear()
@@ -1889,7 +1862,7 @@ class StrixTUIApp(App):  # type: ignore[misc]
                 webbrowser.open(self._viewer_url)
             return
         try:
-            from strix.interface.viewer.server import authorized_url, bundle_is_built, serve
+            from strix.viewer.server import authorized_url, bundle_is_built, serve
 
             if not bundle_is_built():
                 self._set_viewer_cta("[#eab308]Viewer UI not built[/]")

strix/interface/tui/renderers/notes_renderer.py

@@ -7,13 +7,6 @@ from .base_renderer import BaseToolRenderer
 from .registry import register_tool_renderer
 
 
-def _author_label(note: dict[str, Any]) -> str:
-    if note.get("by_you"):
-        return "you"
-    agent_name = note.get("agent_name")
-    return str(agent_name).strip() if agent_name else ""
-
-
 @register_tool_renderer
 class CreateNoteRenderer(BaseToolRenderer):
     tool_name: ClassVar[str] = "create_note"
@@ -130,9 +123,6 @@ class ListNotesRenderer(BaseToolRenderer):
                     text.append("\n  - ")
                     text.append(title)
                     text.append(f" ({category})", style="dim")
-                    author = _author_label(note)
-                    if author:
-                        text.append(f" by {author}", style="dim")
 
                     if note_content:
                         text.append("\n    ")
@@ -166,9 +156,6 @@ class GetNoteRenderer(BaseToolRenderer):
             text.append("\n  ")
             text.append(title)
             text.append(f" ({category})", style="dim")
-            author = _author_label(note)
-            if author:
-                text.append(f" by {author}", style="dim")
             if content:
                 text.append("\n  ")
                 text.append(content, style="dim")

strix/interface/tui/renderers/reporting_renderer.py

@@ -431,117 +431,3 @@ class CreateDependencyReportRenderer(BaseToolRenderer):
 
         css_classes = cls.get_css_classes("completed")
         return Static(padded, classes=css_classes)
-
-
-_LIST_SEVERITY_COLORS = {
-    "critical": "#dc2626",
-    "high": "#ea580c",
-    "medium": "#d97706",
-    "low": "#65a30d",
-    "info": "#0284c7",
-    "none": "#6b7280",
-}
-
-
-def _severity_style(severity: Any) -> str:
-    return _LIST_SEVERITY_COLORS.get(str(severity or "").lower(), "#d97706")
-
-
-def _author_label(report: dict[str, Any]) -> str:
-    if report.get("by_you"):
-        return "you"
-    agent_name = report.get("agent_name")
-    return str(agent_name).strip() if agent_name else ""
-
-
-@register_tool_renderer
-class ListReportsRenderer(BaseToolRenderer):
-    tool_name: ClassVar[str] = "list_reports"
-    css_classes: ClassVar[list[str]] = ["tool-call", "reporting-tool"]
-
-    @classmethod
-    def render(cls, tool_data: dict[str, Any]) -> Static:
-        result = _coerce_dict(tool_data.get("result"))
-
-        text = Text()
-        text.append("◆ ", style="#ef4444")
-        text.append("reports", style="dim")
-
-        if isinstance(tool_data.get("result"), str) and str(tool_data["result"]).strip():
-            text.append("\n  ")
-            text.append(str(tool_data["result"]).strip(), style="dim")
-        elif result.get("success"):
-            total = result.get("total_count", 0)
-            reports = _coerce_list_of_dicts(result.get("reports"))
-            counts = _coerce_dict(result.get("severity_counts"))
-
-            text.append(f" ({total})", style="dim")
-            for sev, count in counts.items():
-                text.append("  ")
-                text.append(f"{sev} {count}", style=_severity_style(sev))
-
-            if not reports:
-                text.append("\n  ")
-                text.append("No reports filed yet", style="dim")
-            else:
-                for report in reports:
-                    rid = str(report.get("id", "")).strip()
-                    title = str(report.get("title", "")).strip() or "(untitled)"
-                    severity = str(report.get("severity", "")).strip()
-                    text.append("\n  - ")
-                    if severity:
-                        text.append(severity.upper(), style=f"bold {_severity_style(severity)}")
-                        text.append(" ")
-                    if rid:
-                        text.append(f"{rid} ", style="dim")
-                    text.append(title)
-                    author = _author_label(report)
-                    if author:
-                        text.append(f" ({author})", style="dim")
-        else:
-            text.append("\n  ")
-            text.append("Loading...", style="dim")
-
-        css_classes = cls.get_css_classes("completed")
-        return Static(text, classes=css_classes)
-
-
-@register_tool_renderer
-class GetReportRenderer(BaseToolRenderer):
-    tool_name: ClassVar[str] = "get_report"
-    css_classes: ClassVar[list[str]] = ["tool-call", "reporting-tool"]
-
-    @classmethod
-    def render(cls, tool_data: dict[str, Any]) -> Static:
-        result = _coerce_dict(tool_data.get("result"))
-
-        text = Text()
-        text.append("◆ ", style="#ef4444")
-        text.append("report read", style="dim")
-
-        report = _coerce_dict(result.get("report")) if result.get("success") else {}
-        if report:
-            rid = str(report.get("id", "")).strip()
-            title = str(report.get("title", "")).strip() or "(untitled)"
-            severity = str(report.get("severity", "")).strip()
-            text.append("\n  ")
-            if severity:
-                text.append(severity.upper(), style=f"bold {_severity_style(severity)}")
-                text.append(" ")
-            if rid:
-                text.append(f"{rid} ", style="dim")
-            text.append(title)
-            author = _author_label(report)
-            if author:
-                text.append(f" ({author})", style="dim")
-            target = str(report.get("target", "")).strip()
-            if target:
-                text.append("\n  ")
-                text.append(target, style="dim")
-        else:
-            text.append("\n  ")
-            detail = result.get("error") if result.get("success") is False else None
-            text.append(str(detail) if detail else "Loading...", style="dim")
-
-        css_classes = cls.get_css_classes("completed")
-        return Static(text, classes=css_classes)

strix/interface/update_check.py

@@ -271,13 +271,7 @@ def _release_target() -> str | None:
     if os_name is None:
         return None
     target = f"{os_name}-{arch}"
-    supported = {
-        "linux-x86_64",
-        "linux-arm64",
-        "macos-x86_64",
-        "macos-arm64",
-        "windows-x86_64",
-    }
+    supported = {"linux-x86_64", "macos-x86_64", "macos-arm64", "windows-x86_64"}
     return target if target in supported else None
 
 

strix/interface/viewer/frontend/src/components/live/tool-renderers/ReportListRenderer.tsx

@@ -1,121 +0,0 @@
-"use client";
-
-import type { ToolRendererProps } from "@/types/events";
-import { TruncatedText } from "./ToolCard";
-import Markdown from "./Markdown";
-
-const SEVERITY_COLORS: Record<string, string> = {
-  critical: "text-red-400", high: "text-orange-400", medium: "text-yellow-400",
-  low: "text-blue-400", info: "text-cyan-400", none: "text-[#888]",
-};
-
-interface ReportEntry {
-  id?: string;
-  title?: string;
-  severity?: string;
-  cvss?: number;
-  cve?: string;
-  cwe?: string;
-  target?: string;
-  endpoint?: string;
-  method?: string;
-  description_preview?: string;
-  description?: string;
-  agent_name?: string;
-  by_you?: boolean;
-}
-
-function authorTag(r: ReportEntry) {
-  if (!r.agent_name && !r.by_you) return null;
-  const label = r.by_you ? "you" : r.agent_name;
-  return <span className="text-[#666] text-xs ml-1.5">({label})</span>;
-}
-
-function sevBadge(severity: string | undefined) {
-  const sev = String(severity ?? "").toLowerCase();
-  const color = SEVERITY_COLORS[sev] ?? "text-yellow-400";
-  return <span className={`font-semibold text-[13px] ${color}`}>{sev.toUpperCase() || "—"}</span>;
-}
-
-export default function ReportListRenderer({ toolName, result }: ToolRendererProps) {
-  const res = result as Record<string, unknown> | null;
-  const ok = res != null && typeof res === "object" && res.success === true;
-
-  if (toolName === "get_report") {
-    const report = ok ? (res.report as ReportEntry | undefined) : undefined;
-    return (
-      <div>
-        <span className="text-red-400/80 font-semibold text-sm">report</span>
-        {report ? (
-          <div className="mt-1.5 space-y-2">
-            <div className="flex items-center gap-2 flex-wrap">
-              {sevBadge(report.severity)}
-              {report.cvss != null && <span className="text-[#888] text-[13px]">CVSS {report.cvss}</span>}
-              {report.id && <span className="text-[#555] font-mono text-[13px]">{report.id}</span>}
-              {report.cve && <span className="text-[#888] font-mono text-[13px]">{report.cve}</span>}
-              {report.cwe && <span className="text-[#888] font-mono text-[13px]">{report.cwe}</span>}
-              {(report.agent_name || report.by_you) && (
-                <span className="text-[#666] text-[13px]">{report.by_you ? "you" : report.agent_name}</span>
-              )}
-            </div>
-            {report.title && <div className="text-[15px] text-white/80 font-semibold">{report.title}</div>}
-            {(report.target || report.endpoint) && (
-              <div className="text-[13px] text-[#888] font-mono">
-                {report.target}{report.endpoint ? ` ${report.method ?? ""} ${report.endpoint}` : ""}
-              </div>
-            )}
-            {report.description && <TruncatedText text={report.description} maxLines={20} />}
-          </div>
-        ) : (
-          <div className="mt-1 text-[#555] text-xs">
-            {(res && typeof res === "object" && (res.error as string)) || "Report not found"}
-          </div>
-        )}
-      </div>
-    );
-  }
-
-  // list_reports
-  const rawReports = ok ? res.reports : null;
-  const reports: ReportEntry[] = Array.isArray(rawReports) ? (rawReports as ReportEntry[]) : [];
-  const total = ok && typeof res.total_count === "number" ? (res.total_count as number) : reports.length;
-  const counts = ok && res.severity_counts && typeof res.severity_counts === "object"
-    ? (res.severity_counts as Record<string, number>)
-    : {};
-  const countEntries = Object.entries(counts);
-
-  return (
-    <div>
-      <div className="flex items-center gap-2 flex-wrap">
-        <span className="text-red-400/80 font-semibold text-sm">reports</span>
-        <span className="text-[#555] text-[13px]">({total})</span>
-        {countEntries.map(([sev, n]) => (
-          <span key={sev} className="text-[13px]">
-            {sevBadge(sev)}<span className="text-[#888] ml-0.5">{n}</span>
-          </span>
-        ))}
-      </div>
-      {reports.length > 0 ? (
-        <div className="mt-1.5 space-y-1">
-          {reports.map((r, i) => (
-            <div key={r.id ?? i} className="text-[13px]">
-              <span className="text-[#555] mr-1">-</span>
-              {sevBadge(r.severity)}
-              {r.id && <span className="text-[#555] font-mono ml-1.5">{r.id}</span>}
-              <span className="text-[#999] ml-1.5">{r.title ?? "(untitled)"}</span>
-              {authorTag(r)}
-              {(r.target || r.endpoint) && (
-                <div className="ml-3 text-[#666] font-mono text-xs">
-                  {r.target}{r.endpoint ? ` ${r.method ?? ""} ${r.endpoint}` : ""}
-                </div>
-              )}
-              {r.description_preview && (
-                <div className="ml-3"><Markdown text={r.description_preview} /></div>
-              )}
-            </div>
-          ))}
-        </div>
-      ) : <div className="mt-1 text-[#555] text-xs">No reports filed yet</div>}
-    </div>
-  );
-}

strix/llm/__init__.py

@@ -1 +0,0 @@
-"""LLM-facing context management: model-aware budgets and history compaction."""

strix/llm/compaction.py

@@ -1,354 +0,0 @@
-"""Provider-agnostic conversation compaction.
-
-When an agent's session grows past the model's usable context window, older
-turns are summarised into a single checkpoint while the most recent turns are
-kept verbatim. This runs for every LiteLLM provider (not just OpenAI), keeps a
-security-focused structured summary, and preserves tool-call/tool-result
-pairing so the trimmed history is still valid provider input.
-"""
-
-from __future__ import annotations
-
-import logging
-from typing import TYPE_CHECKING, Any
-
-import litellm
-from litellm.exceptions import BadRequestError, ContextWindowExceededError
-
-from strix.config import load_settings
-from strix.core.sessions import replace_session_items, session_write_lock
-from strix.llm.context_budget import context_window, count_tokens, output_limit
-
-
-if TYPE_CHECKING:
-    from agents.memory import Session
-
-
-logger = logging.getLogger(__name__)
-
-_CHECKPOINT_TAG = "<conversation-checkpoint>"
-_TOOL_OUTPUT_MAX_CHARS = 2_000
-_MIN_ITEMS_TO_COMPACT = 6
-_HEAD_TRUNCATED_MARKER = "\n\n[... older conversation omitted to fit the summary request ...]\n\n"
-
-
-# Providers that don't type overflow errors (OpenRouter maps every 400 to a
-# plain BadRequestError) leave only the message to go on, so we match it the way
-# LiteLLM's own checker does — but with rate-limit exclusions first, so a
-# throttling 429 is never mistaken for an overflow and sent into compaction.
-_OVERFLOW_EXCLUSIONS = (
-    "rate limit",
-    "too many requests",
-    "throttling",
-    "service unavailable",
-    "quota",
-)
-_OVERFLOW_MARKERS = (
-    "context length",
-    "context window",
-    "context_length_exceeded",
-    "prompt is too long",
-    "input is too long",
-    "input length",
-    "maximum prompt length",
-    "reduce the length of the messages",
-    "too many tokens",
-    "token limit exceeded",
-    "request entity too large",
-)
-
-
-def is_context_overflow(exc: BaseException) -> bool:
-    """Whether ``exc`` is a model context-window-overflow error.
-
-    LiteLLM types most providers' overflow as ContextWindowExceededError, but its
-    OpenRouter branch raises a plain BadRequestError, so for that we fall back to
-    matching the provider message.
-    """
-    if isinstance(exc, ContextWindowExceededError):
-        return True
-    if isinstance(exc, BadRequestError):
-        msg = str(exc).lower()
-        if any(x in msg for x in _OVERFLOW_EXCLUSIONS):
-            return False
-        return any(x in msg for x in _OVERFLOW_MARKERS)
-    return False
-
-
-_SUMMARY_INSTRUCTIONS = """\
-You are compacting the earlier part of an autonomous security-testing agent's \
-conversation so it fits the model context window. Produce a dense, factual \
-record that lets the agent continue with no loss of important state.
-
-This is a security engagement: dropped findings mean lost vulnerabilities. Be \
-EXHAUSTIVE, not concise. Enumerate every distinct item as its own bullet — \
-never merge, deduplicate, generalise, or omit distinct findings, credentials, \
-or dead ends, even if they seem minor or repetitive. If the source mentions \
-five vulnerabilities, list five. Copy exact values verbatim: URLs, endpoints, \
-file paths, parameters, payloads, credentials, tokens, keys, hashes, cracked \
-passwords, software versions, and error messages — never paraphrase or \
-placeholder them. Do not invent anything and do not describe this compaction \
-process.
-
-Return Markdown with exactly these sections:
-
-## Objective
-The overall goal and target scope.
-
-## Vulnerabilities & Findings
-One bullet per DISTINCT vulnerability or finding (SQLi, XSS, SSRF, auth bypass, \
-misconfig, etc.). For each: type, exact location (URL/endpoint/param/file), the \
-verbatim payload or proof, confirmation status, and impact. List them all.
-
-## Credentials & Secrets
-One bullet per credential, secret, API key, token, hash, or cracked password, \
-copied verbatim with where it applies. Write "(none)" only if truly none.
-
-## System & Recon Details
-Architecture, tech stack, versions, discovered endpoints/paths/params, and \
-other weak points worth keeping.
-
-## Work State
-- Completed: what has been verified or finished.
-- Active: what is in progress right now.
-- Blocked: anything stuck and why.
-
-## Failed Attempts & Dead Ends
-One bullet per approach already tried that did not work (including WAF blocks, \
-filtered inputs, non-exploitable leads) so they are not repeated. Write \
-"(none)" only if truly none.
-
-## Next Move
-The concrete next step(s) the agent intended to take.
-
-## Relevant Files
-Files/notes/reports created or modified and their purpose."""
-
-
-def _content_text(content: Any) -> str:
-    if isinstance(content, str):
-        return content
-    if isinstance(content, list):
-        parts: list[str] = []
-        for block in content:
-            if not isinstance(block, dict):
-                continue
-            text = block.get("text")
-            if isinstance(text, str):
-                parts.append(text)
-            elif block.get("type") in {"input_image", "image_url", "output_image"}:
-                parts.append("[image]")
-        return "\n".join(parts)
-    return ""
-
-
-def _truncate(text: str, limit: int) -> str:
-    return text if len(text) <= limit else f"{text[:limit]}\n[truncated]"
-
-
-def _serialize_item(item: Any) -> str:
-    if not isinstance(item, dict):
-        return str(item)
-    item_type = item.get("type")
-    role = item.get("role")
-    if item_type == "function_call":
-        args = _truncate(str(item.get("arguments", "")), _TOOL_OUTPUT_MAX_CHARS)
-        return f"[tool_call {item.get('name', '?')}] {args}"
-    if item_type == "function_call_output":
-        output = item.get("output")
-        text = output if isinstance(output, str) else _content_text(output)
-        return f"[tool_result] {_truncate(text, _TOOL_OUTPUT_MAX_CHARS)}"
-    if item_type == "reasoning":
-        return ""
-    if role or item_type == "message":
-        return f"[{role or 'assistant'}] {_content_text(item.get('content'))}".strip()
-    return ""
-
-
-def _serialize_items(items: list[Any]) -> str:
-    return "\n".join(s for s in (_serialize_item(item) for item in items) if s)
-
-
-def _is_tool_call(item: Any) -> bool:
-    return isinstance(item, dict) and item.get("type") == "function_call"
-
-
-def _is_tool_output(item: Any) -> bool:
-    return isinstance(item, dict) and item.get("type") == "function_call_output"
-
-
-def _open_calls_at(items: list[Any]) -> list[int]:
-    """Prefix count of tool calls still awaiting their result at each index;
-    a split is only safe where this is zero."""
-    balance = [0] * (len(items) + 1)
-    for i, item in enumerate(items):
-        delta = 1 if _is_tool_call(item) else -1 if _is_tool_output(item) else 0
-        balance[i + 1] = max(0, balance[i] + delta)
-    return balance
-
-
-def _select_split(model: str, items: list[Any], keep_tokens: int) -> int:
-    """Index where the kept-verbatim recent tail begins: walk newest→oldest to
-    ``keep_tokens``, then snap to a point with no tool call left open."""
-    total = 0
-    split = len(items)
-    for i in range(len(items) - 1, -1, -1):
-        total += count_tokens(model, _serialize_item(items[i]))
-        if total > keep_tokens:
-            break
-        split = i
-    open_calls = _open_calls_at(items)
-    while split > 0 and open_calls[split] != 0:
-        split -= 1
-    return split
-
-
-def _previous_summary(head: list[Any]) -> str | None:
-    for item in head:
-        if isinstance(item, dict) and item.get("role") == "user":
-            text = _content_text(item.get("content"))
-            if text.startswith(_CHECKPOINT_TAG):
-                return text
-    return None
-
-
-def _fit_to_tokens(model: str, text: str, max_tokens: int) -> str:
-    """Head+tail-truncate ``text`` to ``max_tokens``, keeping start and end."""
-    if count_tokens(model, text) <= max_tokens:
-        return text
-    # Rough char budget (~4x tokens), then tighten by real token count.
-    budget_chars = max_tokens * 4
-    head_chars = budget_chars // 2
-    tail_chars = budget_chars - head_chars
-    candidate = text[:head_chars] + _HEAD_TRUNCATED_MARKER + text[len(text) - tail_chars :]
-    while count_tokens(model, candidate) > max_tokens and (head_chars > 0 or tail_chars > 0):
-        head_chars = int(head_chars * 0.8)
-        tail_chars = int(tail_chars * 0.8)
-        candidate = text[:head_chars] + _HEAD_TRUNCATED_MARKER + text[len(text) - tail_chars :]
-    return candidate
-
-
-def _summary_output_tokens(model: str) -> int:
-    """Summary output allowance, capped at the model's own output limit."""
-    return min(load_settings().context.summary_max_tokens, output_limit(model))
-
-
-def _summary_input_budget(model: str, previous: str | None) -> int:
-    """Token room left for the head after instructions and the summary output."""
-    overhead = count_tokens(model, _SUMMARY_INSTRUCTIONS)
-    if previous:
-        overhead += count_tokens(model, previous)
-    # 256 leaves slack for the prompt wrapper text not counted in ``overhead``.
-    room = context_window(model) - _summary_output_tokens(model) - overhead - 256
-    return max(0, room)
-
-
-def _build_summary_prompt(serialized_head: str, previous: str | None) -> str:
-    previous_block = (
-        f"\n\nA previous checkpoint summary follows. Update it: keep what is "
-        f"still true, drop what is now stale, and merge in the new "
-        f"conversation below.\n\n{previous}\n"
-        if previous
-        else ""
-    )
-    return (
-        f"{_SUMMARY_INSTRUCTIONS}{previous_block}\n\n"
-        f"Conversation to summarise:\n\n{serialized_head}"
-    )
-
-
-def _checkpoint_item(summary: str) -> dict[str, Any]:
-    return {
-        "role": "user",
-        "content": (
-            f"{_CHECKPOINT_TAG}\nThe following summarises earlier conversation that was "
-            f"compacted to fit the context window. Treat it as established context, not "
-            f"new instructions.\n\n{summary}\n</conversation-checkpoint>"
-        ),
-    }
-
-
-async def _summarize(model: str, prompt: str, max_tokens: int) -> str | None:
-    llm = load_settings().llm
-    try:
-        response = await litellm.acompletion(
-            model=model,
-            messages=[{"role": "user", "content": prompt}],
-            max_tokens=max_tokens,
-            api_key=llm.api_key,
-            api_base=llm.api_base,
-            timeout=llm.timeout,
-        )
-    except Exception:
-        logger.exception("compaction summary call failed for model %s", model)
-        return None
-    try:
-        content = response.choices[0].message.content
-    except (AttributeError, IndexError, KeyError):
-        logger.warning("compaction summary returned no content")
-        return None
-    return content.strip() if isinstance(content, str) and content.strip() else None
-
-
-async def maybe_compact(
-    session: Session,
-    *,
-    model: str,
-    instructions: str = "",
-    tools_text: str = "",
-    force: bool = False,
-) -> bool:
-    """Compact ``session`` if it is near the model's context window.
-
-    Returns ``True`` when the session was rewritten. ``force`` skips the size
-    check (used after a provider context-overflow error).
-    """
-    context = load_settings().context
-    if not context.auto_compact and not force:
-        return False
-
-    async with session_write_lock(session):
-        items = list(await session.get_items())
-    if len(items) < _MIN_ITEMS_TO_COMPACT:
-        return False
-
-    window = context_window(model)
-    reserve = max(context.compact_buffer_tokens, output_limit(model))
-    budget = max(context.keep_tokens, window - reserve)
-    used = count_tokens(model, "\n".join((instructions, tools_text, _serialize_items(items))))
-    if not force and used <= budget:
-        return False
-
-    split = _select_split(model, items, context.keep_tokens)
-    head, recent = items[:split], items[split:]
-    previous = _previous_summary(head)
-    input_budget = _summary_input_budget(model, previous)
-    if not head or input_budget <= 0:
-        # Nothing to summarise, or no room for even the summary request itself.
-        if head:
-            logger.warning(
-                "skipping compaction for %s: no room to summarise within its context window", model
-            )
-        return False
-
-    serialized_head = _fit_to_tokens(model, _serialize_items(head), input_budget)
-    summary = await _summarize(
-        model,
-        _build_summary_prompt(serialized_head, previous),
-        _summary_output_tokens(model),
-    )
-    if summary is None:
-        return False
-
-    new_items = [_checkpoint_item(summary), *recent]
-    rewritten = await replace_session_items(session, new_items, expected_len=len(items))
-    if rewritten:
-        logger.info(
-            "compacted %s: %d items (~%d tok) -> %d items (summary + %d recent)",
-            model,
-            len(items),
-            used,
-            len(new_items),
-            len(recent),
-        )
-    return rewritten

strix/llm/context_budget.py

@@ -1,76 +0,0 @@
-"""Model-aware token budgets, resolved from LiteLLM model metadata with a
-large configurable fallback for models LiteLLM doesn't map.
-"""
-
-from __future__ import annotations
-
-import logging
-from functools import lru_cache
-from typing import Any
-
-import litellm
-
-from strix.config import load_settings
-
-
-logger = logging.getLogger(__name__)
-
-# LiteLLM keys models without the routing prefix users type (``openai/``,
-# ``litellm/``, ``ollama/`` ...). Strip a leading provider segment on lookup.
-_STRIPPABLE_PREFIXES = ("openai/", "litellm/", "any-llm/", "ollama/", "ollama_chat/")
-
-_DEFAULT_OUTPUT_TOKENS = 8_192
-
-
-def _lookup_key(model: str) -> str:
-    for prefix in _STRIPPABLE_PREFIXES:
-        if model.startswith(prefix):
-            return model[len(prefix) :]
-    return model
-
-
-def _safe_get_model_info(model: str) -> dict[str, Any] | None:
-    try:
-        return dict(litellm.get_model_info(model))
-    except Exception:  # noqa: BLE001 - unmapped models raise; caller falls back.
-        return None
-
-
-@lru_cache(maxsize=128)
-def _model_info(model: str) -> dict[str, int]:
-    for candidate in (model, _lookup_key(model)):
-        info = _safe_get_model_info(candidate)
-        if info is not None:
-            return {
-                "max_input_tokens": int(
-                    info.get("max_input_tokens") or info.get("max_tokens") or 0
-                ),
-                "max_output_tokens": int(info.get("max_output_tokens") or 0),
-            }
-    logger.debug("No LiteLLM model info for %r; using configured fallbacks", model)
-    return {"max_input_tokens": 0, "max_output_tokens": 0}
-
-
-def context_window(model: str) -> int:
-    """Input token capacity for ``model`` (configured fallback when unmapped)."""
-    resolved = _model_info(model)["max_input_tokens"]
-    return resolved or load_settings().context.fallback_context_tokens
-
-
-def output_limit(model: str) -> int:
-    """Max output tokens for ``model`` (a conservative default when unmapped)."""
-    return _model_info(model)["max_output_tokens"] or _DEFAULT_OUTPUT_TOKENS
-
-
-def count_tokens(model: str, text: str) -> int:
-    """Token count for ``text`` under ``model``.
-
-    Falls back to UTF-8 byte length (a guaranteed upper bound) when LiteLLM
-    can't count, so budget checks stay conservative.
-    """
-    if not text:
-        return 0
-    try:
-        return int(litellm.token_counter(model=_lookup_key(model), text=text))
-    except Exception:  # noqa: BLE001 - tokenizer may be unavailable for some models.
-        return len(text.encode("utf-8"))

strix/skills/reconnaissance/asset_discovery.md

@@ -54,17 +54,18 @@ CT logs record nearly every publicly-trusted certificate. Query by domain (match
 
 ## Recommended Tooling
 
-These tools are available in the sandbox and are pipeline-friendly with JSON output:
+Prefer the projectdiscovery suite (already available in the sandbox and pipeline-friendly with JSON output):
 
 - **`subfinder`** — passive subdomain aggregation across many sources incl. CT: `subfinder -d example.com -all -recursive -silent -oJ -o subs.jsonl`
-- **`httpx`** — live probing plus cert/SAN grab in one pass: `httpx -l hosts.txt -tls-grab -json` (see methodology).
+- **`tlsx`** — TLS/cert data at scale; grab SANs and issuer/org to pivot: `tlsx -l hosts.txt -san -cn -tls-version -json -o tls.jsonl`
+- **`uncover`** — query Shodan/Censys/Fofa/Quake/crt.sh engines from one CLI: `uncover -q 'ssl:"Example Inc"' -e shodan,censys,fofa -json`
+- **`asnmap`** — org/domain/ASN → CIDR ranges: `asnmap -d example.com -json` / `asnmap -org "Example Inc"`
+- **`mapcidr`** — expand/aggregate CIDRs into host lists for probing: `mapcidr -cidr 192.0.2.0/24 -o hosts.txt`
+- **`dnsx`** — fast resolution, PTR, and wildcard filtering: `dnsx -l names.txt -a -aaaa -cname -ptr -resp -json -o dns.jsonl`
+- **`httpx`** — live probing + cert grab in one pass (see methodology).
 - **`naabu`** — port sweep for non-HTTP services: `naabu -list hosts.txt -top-ports 100 -verify -silent`
-- **`curl` + `jq`** — direct **crt.sh** JSON queries for CT (no key needed) and other index APIs.
-- **`openssl s_client`** — active read of a live host's cert to extract SANs/CN.
-- **`dig`** / **`nslookup`** — forward/reverse (PTR) resolution and CNAME chains.
-- **`whois`** — ASN/netblock lookups (e.g. `whois -h whois.cymru.com`).
 
-Cross-source results — CT + passive DNS + `subfinder` together beat any single source. If you need a tool that is not installed, install it into the sandbox at runtime.
+Also useful: **`amass`** (`amass intel`/`enum` for ASN, cert, and passive sources), **`cero`** (bulk SAN extraction from IPs/ranges), and direct **crt.sh** JSON queries when no keys are configured. Cross-source results — CT + passive DNS + `subfinder` together beat any single source.
 
 ## Key Techniques
 
@@ -74,7 +75,7 @@ Every new name, PTR result, CNAME target, and cert SAN becomes a fresh seed. Loo
 
 ### Cert-Fingerprint Pivoting
 
-Search Censys/Shodan by a cert's `fingerprint_sha256` to find every other host presenting the same certificate — the strongest cross-asset link for tying acquisitions and shadow infra to the target.
+Search Censys/Shodan (or `uncover`) by a cert's `fingerprint_sha256` to find every other host presenting the same certificate — the strongest cross-asset link for tying acquisitions and shadow infra to the target.
 
 ### Naming-Convention Inference
 
@@ -82,11 +83,11 @@ Wildcard SANs and observed hostnames expose the org's naming scheme; generate ta
 
 ### IP-First Discovery
 
-For ASN-owned ranges, sweep IPs directly with `naabu`/`httpx` and read served certs (`httpx -tls-grab`, or `openssl s_client`) to find services that have no DNS name at all.
+For ASN-owned ranges, sweep IPs directly with `naabu`/`httpx` and read served certs (`tlsx`) to find services that have no DNS name at all.
 
 ## Advanced Techniques
 
-- **Active SAN harvesting** across whole ranges with `httpx -tls-grab` (or `openssl s_client`) recovers internal hostnames never logged to public CT.
+- **Active SAN harvesting** across whole ranges with `tlsx`/`cero` recovers internal hostnames never logged to public CT.
 - **Favicon and response hashing** (`httpx -favicon`, hash pivots in Shodan) clusters instances of the same app across unrelated hostnames.
 - **Vhost differentials**: probe a single IP with multiple `Host:` values to unmask co-located apps behind one address.
 - **Historical CT/DNS diffing** highlights recently issued certs and newly appearing hosts — high-signal for fresh or misconfigured deployments.
@@ -107,11 +108,11 @@ For ASN-owned ranges, sweep IPs directly with `naabu`/`httpx` and read served ce
 ## Testing Methodology
 
 1. **Seed** - domains, org/legal names, known IPs, email domains, code-host org
-2. **Certificate transparency** - pull all logged certs per seed domain and org name (crt.sh, Censys/Shodan)
-3. **SAN/CN extraction** - parse every Subject CN and SAN with `httpx -tls-grab` (or `openssl s_client`); each new name is a new seed
-4. **Passive DNS** - resolve forward and reverse with `dig`; harvest historical records
-5. **ASN/IP mapping** - `whois` the netblock/ASN to expand owned ranges, then sweep for live hosts
-6. **Active TLS pivot** - `httpx -tls-grab` on live IPs/ports to grab SANs missing from public CT
+2. **Certificate transparency** - pull all logged certs per seed domain and org name (crt.sh, `uncover`)
+3. **SAN/CN extraction** - parse every Subject CN and SAN with `tlsx`; each new name is a new seed
+4. **Passive DNS** - resolve forward and reverse with `dnsx`; harvest historical records
+5. **ASN/IP mapping** - `asnmap` → `mapcidr` to expand owned ranges, then sweep for live hosts
+6. **Active TLS pivot** - `tlsx`/`cero` on live IPs/ports to grab SANs missing from public CT
 7. **Consolidate & probe** - dedupe, `httpx` probe, classify, and route to specialists
 
 ## Validation
@@ -138,13 +139,13 @@ For ASN-owned ranges, sweep IPs directly with `naabu`/`httpx` and read served ce
 ## Pro Tips
 
 1. Loop the pipeline — every SAN, PTR, and CNAME target is a new seed until the set converges.
-2. crt.sh is the cheapest high-yield source (no key); Censys/Shodan add cert-fingerprint and vhost pivoting when keys exist.
-3. Always cert-grab live hosts with `httpx -tls-grab` (or `openssl s_client`) — active SANs catch internal hostnames never sent to public CT.
+2. crt.sh is the cheapest high-yield source (no key); Censys/Shodan via `uncover` add cert-fingerprint and vhost pivoting when keys exist.
+3. Always cert-grab live hosts with `tlsx` — active SANs catch internal hostnames never sent to public CT.
 4. Internal-looking SANs (`*.internal`, `*.svc.cluster.local`, staging names) are the highest-signal leads.
 5. Wildcard SANs reveal naming conventions — seed targeted guesses instead of blind brute force.
 6. Cluster by function, not product name, so the workflow generalizes to any exposed service.
-7. Keep JSON output throughout so stages chain cleanly (`subfinder` → `dig` → `httpx` → `naabu`).
+7. Keep JSON output throughout so stages chain cleanly (`subfinder` → `dnsx` → `httpx` → `naabu`).
 
 ## Summary
 
-Broad passive discovery — CT + TLS SAN pivoting + passive DNS + ASN/IP mapping, looped until convergence — finds the assets brute force misses, especially internal-named and forgotten services leaked through certificates. Build the inventory with `subfinder`, `httpx`, `naabu`, and CT/DNS/cert queries, probe and classify it generically, then route each interesting asset to the specialist skill for its class.
+Broad passive discovery — CT + TLS SAN pivoting + passive DNS + ASN/IP mapping, looped until convergence — finds the assets brute force misses, especially internal-named and forgotten services leaked through certificates. Build the inventory with the projectdiscovery suite, probe and classify it generically, then route each interesting asset to the specialist skill for its class.

strix/skills/vulnerabilities/subdomain_takeover.md

@@ -152,7 +152,7 @@ TLS clues: certificate CN/SAN referencing provider default host instead of the c
 
 ## Pro Tips
 
-1. Build a pipeline: enumerate (subfinder) → resolve (dig) → probe (httpx) → fingerprint (nuclei/custom) → verify claims
+1. Build a pipeline: enumerate (subfinder/amass) → resolve (dnsx) → probe (httpx) → fingerprint (nuclei/custom) → verify claims
 2. Maintain a current fingerprint corpus; provider messages change frequently
 3. Prefer minimal PoCs: static "ownership proof" page and, where allowed, DV cert issuance
 4. Monitor CT for unexpected certs on your subdomains

strix/skills/vulnerabilities/weak_password_detection.md

@@ -1,6 +1,6 @@
 ---
 name: weak-password-detection
-description: Weak password detection, credential stuffing, and brute-force testing using common passwords, system-generated credentials, and HTTP fuzzing / NSE brute-force tooling
+description: Weak password detection, credential stuffing, and brute-force testing using common passwords, system-generated credentials, and tooling like Hydra
 ---
 
 # Weak Password Detection / Credential Brute-Force
@@ -98,7 +98,7 @@ Weak or default credentials remain one of the most prevalent and high-impact vul
   - Season + year patterns: `Summer2025!`, `Winter2026@`
   - Keyboard walks and leet speak variations
   - Previously breached passwords for the target domain
-- Scrape the target site to build a content-derived wordlist (e.g. a small custom Python crawler that harvests unique words)
+- Cewl: `cewl -d 3 -m 5 -w custom.txt https://target.com` to generate from website content
 
 ### Credential Stuffing Workflows
 
@@ -123,25 +123,37 @@ Weak or default credentials remain one of the most prevalent and high-impact vul
 
 ### Service-Level Brute-Force
 
-- HTTP login endpoints: `ffuf` or custom scripts (see Tooling)
-- SSH/FTP/SMB/Telnet and other services: `nmap` NSE `*-brute` scripts, e.g. `nmap -p 22 --script ssh-brute --script-args userdb=users.txt,passdb=passwords.txt target.com`
-- Databases (MySQL, PostgreSQL, MongoDB, Redis): weak/default credentials via the matching NSE brute script (`mysql-brute`, `pgsql-brute`, `mongodb-brute`, `redis-brute`) or a custom client script
-- Any protocol lacking a ready script: custom Python
+- SSH: `hydra -l admin -P passwords.txt ssh://target.com`
+- FTP: `hydra -L users.txt -P passwords.txt ftp://target.com`
+- RDP: `hydra -l administrator -P passwords.txt rdp://target.com`
+- SMB: `hydra -L users.txt -P passwords.txt smb://target.com`
+- Database: MySQL, PostgreSQL, MongoDB, Redis with weak credentials
+- API endpoints: `ffuf` or custom scripts for HTTP-based brute-force
 
 ## Tooling
 
-### ffuf (primary for web logins)
+### Hydra (Primary Tool)
+
+- HTTP POST form brute-force:
+  `hydra -l admin -P /usr/share/wordlists/rockyou.txt target.com http-post-form "/login:username=^USER^&password=^PASS^:Invalid credentials"`
+- Basic Auth:
+  `hydra -L users.txt -P passwords.txt target.com http-get -s 8080 /admin`
+- SSH:
+  `hydra -l root -P passwords.txt -t 4 ssh://target.com`
+- FTP:
+  `hydra -L users.txt -P passwords.txt ftp://target.com`
+- Custom headers and cookies:
+  `hydra ... http-post-form "/api/login:json={\"user\":\"^USER^\",\"pass\":\"^PASS^\"}:F=401"`
+
+### ffuf (HTTP Fuzzing)
 
 - Login brute-force with multiple users and passwords:
   `ffuf -w users.txt:USER -w passwords.txt:PASS -u https://target.com/login -X POST -d "username=USER&password=PASS" -fr "Invalid"`
-- JSON body / custom headers via `-H` and a JSON `-d` payload
 - Filter by response size, status code, or regex to identify successes
 
-### nmap NSE (service brute-force)
+### Patator (Versatile Brute-Force)
 
-- `*-brute` scripts cover many non-HTTP services:
-  `nmap -p 22 --script ssh-brute --script-args userdb=users.txt,passdb=passwords.txt target.com`
-- Available scripts include `ssh-brute`, `ftp-brute`, `smb-brute`, `telnet-brute`, `mysql-brute`, `pgsql-brute`, `mongodb-brute`, `redis-brute`, `http-brute`, `http-form-brute`.
+- `patator http_fuzz url=https://target.com/login method=POST body='username=FILE0&password=FILE1' 0=user.txt 1=pass.txt -x ignore:fgrep='Invalid'`
 
 ### Custom Python Scripts
 
@@ -151,10 +163,10 @@ Weak or default credentials remain one of the most prevalent and high-impact vul
 
 ### Wordlists
 
-No password wordlists ship in the sandbox by default — download what you need into `/home/pentester/tools/wordlists` at runtime:
-- Common passwords (e.g. `rockyou.txt`) from its upstream source
-- SecLists `Passwords/` and `Passwords/Default-Credentials/` (vendor defaults) from https://github.com/danielmiessler/SecLists
-- Custom lists from target-specific scraping
+- `/usr/share/wordlists/rockyou.txt` (common passwords)
+- `/usr/share/seclists/Passwords/` (organized by category)
+- `/usr/share/seclists/Passwords/Default-Credentials/` (vendor defaults)
+- Custom lists from Cewl, CeWL, or target-specific scraping
 - Breach compilation subsets filtered by target relevance
 
 ## Validation
@@ -192,7 +204,7 @@ No password wordlists ship in the sandbox by default — download what you need
 6. Check for concurrent session limits; successful logins may kick out legitimate users
 7. GraphQL batching can test multiple credentials in a single request, bypassing per-request limits
 8. Document the password policy and recommend minimum standards (length, complexity, breach checking)
-9. For web logins prefer `ffuf`; for other services use `nmap` NSE `*-brute` scripts or custom scripts with equivalent logic
+9. When Hydra is unavailable, use ffuf or custom scripts with equivalent logic
 10. Combine with MFA testing: weak passwords plus missing MFA is a critical finding
 
 ## Summary

strix/tools/finish/tool.py

@@ -116,16 +116,12 @@ async def finish_scan(
        / ``crashed`` / ``stopped`` agents are safe to leave behind.
        Calling ``finish_scan`` while children are alive orphans their
        work and produces an incomplete report.
-    2. It's a good idea to call ``list_reports`` before finishing to
-       review every finding filed in this scan (use ``get_report`` for
-       full detail on any of them) so your ``executive_summary`` /
-       ``technical_analysis`` are grounded in what was actually reported
-       — don't invent or omit findings. All vulnerabilities you found are
-       filed via ``create_vulnerability_report`` — or, for known-CVE
-       dependency findings, ``create_dependency_report`` (un-reported
-       findings are not tracked and not credited). A dependency CVE
-       already filed via ``create_dependency_report`` counts as reported;
-       it does NOT need re-filing here and does NOT block finishing.
+    2. All vulnerabilities you found are filed via
+       ``create_vulnerability_report`` — or, for known-CVE dependency
+       findings, ``create_dependency_report`` (un-reported findings are
+       not tracked and not credited). A dependency CVE already filed via
+       ``create_dependency_report`` counts as reported; it does NOT need
+       re-filing here and does NOT block finishing.
     3. Don't double-report — one report per distinct vulnerability.
     4. **Attack-chaining gate.** Do NOT finish until you have genuinely
        considered chaining the confirmed findings into higher-impact,
@@ -253,8 +249,6 @@ async def finish_scan(
     parent_id = inner.get("parent_id")
     if coordinator is not None and parent_id is None and me is not None:
         active_agents = await coordinator.active_agents_except(me)
-        if active_agents and coordinator.reserve_stopped:
-            active_agents = []
     else:
         active_agents = []
 

strix/tools/notes/tools.py

@@ -27,21 +27,6 @@ _NOTE_ID_GENERATION_ATTEMPTS = 1024
 _notes_path: Path | None = None
 
 
-def _caller_identity(ctx: RunContextWrapper) -> tuple[str | None, str | None]:
-    """Return the (agent_id, agent_name) of the agent invoking this tool."""
-    inner = ctx.context if isinstance(ctx.context, dict) else {}
-    raw_agent_id = inner.get("agent_id")
-    agent_id = raw_agent_id if isinstance(raw_agent_id, str) else None
-    agent_name: str | None = None
-    coordinator = inner.get("coordinator")
-    if agent_id is not None and coordinator is not None:
-        names = getattr(coordinator, "names", {})
-        if isinstance(names, dict):
-            raw_agent_name = names.get(agent_id)
-            agent_name = raw_agent_name if isinstance(raw_agent_name, str) else None
-    return agent_id, agent_name
-
-
 def _generate_note_id() -> str | None:
     for _ in range(_NOTE_ID_GENERATION_ATTEMPTS):
         note_id = uuid.uuid4().hex[:6]
@@ -132,26 +117,10 @@ def _filter_notes(
     return filtered
 
 
-def _mark_authorship(
-    entry: dict[str, Any], note: dict[str, Any], caller_agent_id: str | None
-) -> dict[str, Any]:
-    """Attach the note's author and flag whether the caller wrote it."""
-    agent_name = note.get("agent_name")
-    if agent_name:
-        entry["agent_name"] = agent_name
-    agent_id = note.get("agent_id")
-    if agent_id:
-        entry["agent_id"] = agent_id
-    if caller_agent_id is not None and agent_id == caller_agent_id:
-        entry["by_you"] = True
-    return entry
-
-
 def _to_note_listing_entry(
     note: dict[str, Any],
     *,
     include_content: bool = False,
-    caller_agent_id: str | None = None,
 ) -> dict[str, Any]:
     entry = {
         "note_id": note.get("note_id"),
@@ -169,7 +138,7 @@ def _to_note_listing_entry(
             entry["content_preview"] = f"{content[:_DEFAULT_CONTENT_PREVIEW_CHARS].rstrip()}..."
         else:
             entry["content_preview"] = content
-    return _mark_authorship(entry, note, caller_agent_id)
+    return entry
 
 
 def _create_note_impl(
@@ -177,8 +146,6 @@ def _create_note_impl(
     content: str,
     category: str = "general",
     tags: list[str] | None = None,
-    agent_id: str | None = None,
-    agent_name: str | None = None,
 ) -> dict[str, Any]:
     with _notes_lock:
         try:
@@ -212,10 +179,6 @@ def _create_note_impl(
                 "created_at": timestamp,
                 "updated_at": timestamp,
             }
-            if agent_id:
-                note["agent_id"] = agent_id
-            if agent_name:
-                note["agent_name"] = agent_name
             _notes_storage[note_id] = note
         except (ValueError, TypeError) as e:
             return {"success": False, "error": f"Failed to create note: {e}", "note_id": None}
@@ -234,17 +197,11 @@ def _list_notes_impl(
     tags: list[str] | None = None,
     search: str | None = None,
     include_content: bool = False,
-    caller_agent_id: str | None = None,
 ) -> dict[str, Any]:
     with _notes_lock:
         try:
             filtered = _filter_notes(category=category, tags=tags, search_query=search)
-            notes = [
-                _to_note_listing_entry(
-                    n, include_content=include_content, caller_agent_id=caller_agent_id
-                )
-                for n in filtered
-            ]
+            notes = [_to_note_listing_entry(n, include_content=include_content) for n in filtered]
         except (ValueError, TypeError) as e:
             return {
                 "success": False,
@@ -261,7 +218,7 @@ def _list_notes_impl(
         }
 
 
-def _get_note_impl(note_id: str, caller_agent_id: str | None = None) -> dict[str, Any]:
+def _get_note_impl(note_id: str) -> dict[str, Any]:
     with _notes_lock:
         try:
             if not note_id or not note_id.strip():
@@ -275,7 +232,6 @@ def _get_note_impl(note_id: str, caller_agent_id: str | None = None) -> dict[str
                 }
             note_with_id = note.copy()
             note_with_id["note_id"] = note_id
-            _mark_authorship(note_with_id, note, caller_agent_id)
         except (ValueError, TypeError) as e:
             return {"success": False, "error": f"Failed to get note: {e}", "note": None}
         else:
@@ -348,9 +304,7 @@ async def create_note(
 
     Notes are visible to every agent in the same scan for the lifetime
     of the run; they live in-memory only and are cleared when the
-    process exits. Each note records the agent that wrote it, so
-    ``list_notes`` / ``get_note`` show the author (``agent_name``) and
-    flag your own notes with ``by_you``.
+    process exits.
 
     For actionable tasks, use ``todo`` instead — notes are for capturing
     information, todos are for tracking work.
@@ -375,11 +329,8 @@ async def create_note(
         category: One of the categories above. Default ``"general"``.
         tags: Optional free-form tags.
     """
-    agent_id, agent_name = _caller_identity(ctx)
     return json.dumps(
-        await asyncio.to_thread(
-            _create_note_impl, title, content, category, tags, agent_id, agent_name
-        ),
+        await asyncio.to_thread(_create_note_impl, title, content, category, tags),
         ensure_ascii=False,
         default=str,
     )
@@ -404,9 +355,6 @@ async def list_notes(
     when you need to scan many notes; expensive in tokens for large
     notes.
 
-    Each entry also carries the author (``agent_name``) and, for notes
-    you wrote yourself, ``by_you: true``.
-
     Args:
         category: Filter by category.
         tags: Filter to notes that have any of these tags.
@@ -414,7 +362,6 @@ async def list_notes(
         include_content: When False (default) entries have a preview;
             when True the full ``content`` is included.
     """
-    caller_agent_id, _ = _caller_identity(ctx)
     return json.dumps(
         await asyncio.to_thread(
             _list_notes_impl,
@@ -422,7 +369,6 @@ async def list_notes(
             tags=tags,
             search=search,
             include_content=include_content,
-            caller_agent_id=caller_agent_id,
         ),
         ensure_ascii=False,
         default=str,
@@ -436,11 +382,8 @@ async def get_note(ctx: RunContextWrapper, note_id: str) -> str:
     Args:
         note_id: Note id from ``create_note`` or a ``list_notes`` entry.
     """
-    caller_agent_id, _ = _caller_identity(ctx)
     return json.dumps(
-        await asyncio.to_thread(_get_note_impl, note_id, caller_agent_id),
-        ensure_ascii=False,
-        default=str,
+        await asyncio.to_thread(_get_note_impl, note_id), ensure_ascii=False, default=str
     )
 
 

strix/tools/output_store.py

@@ -1,47 +1,13 @@
 """Bound oversized tool results before they enter agent history.
 
-Oversized results are spilled into the sandbox at
-``/workspace/.strix/tool-output/<id>.txt``; the agent sees a head + tail slice
-plus the path and reads the rest back with its own file tools. The spill writer
-is injected by the runner via :func:`configure_spill_writer`.
+Keeps a head + tail slice and drops the middle, replacing it with a notice of
+how much was removed.
 """
 
 from __future__ import annotations
 
-import logging
-import uuid
-from typing import TYPE_CHECKING
-
-
-if TYPE_CHECKING:
-    from collections.abc import Awaitable, Callable
-
-
-logger = logging.getLogger(__name__)
 
 _TRUNCATION_NOTICE = "[... {lines} lines ({bytes} bytes) truncated ...]"
-_WORKSPACE_SPILL_NOTICE = (
-    "[... {lines} lines ({bytes} bytes) truncated — full output saved to {path} "
-    "in the sandbox; read it with exec_command (e.g. `sed -n`, `grep`, `cat`) ...]"
-)
-
-WORKSPACE_SPILL_DIR = "/workspace/.strix/tool-output"
-
-# Longest possible workspace path, used only to reserve notice bytes.
-_SAMPLE_WORKSPACE_PATH = f"{WORKSPACE_SPILL_DIR}/{'0' * 32}.txt"
-
-if TYPE_CHECKING:
-    SpillWriter = Callable[[str, str], Awaitable[str | None]]
-
-_spill: dict[str, SpillWriter] = {}
-
-
-def configure_spill_writer(writer: SpillWriter | None) -> None:
-    """Install (or clear) the sandbox-workspace spill writer."""
-    if writer is None:
-        _spill.pop("writer", None)
-    else:
-        _spill["writer"] = writer
 
 
 def _byte_len(text: str) -> int:
@@ -73,37 +39,20 @@ def _take_suffix(text: str, max_bytes: int) -> str:
     return "".join(out)
 
 
-def _head_tail(
-    text: str,
-    max_lines: int,
-    max_bytes: int,
-    *,
-    notice_templates: tuple[str, ...] = (_TRUNCATION_NOTICE,),
-) -> tuple[str, str, int, int] | None:
-    """Head/tail slices plus dropped line/byte counts, or ``None`` if small.
+def bound_text(text: str, *, max_lines: int, max_bytes: int) -> str:
+    """Return ``text`` unchanged when small, else a head+tail preview.
 
-    ``max_bytes`` bounds the entire joined result; the largest of
-    ``notice_templates`` (plus separators) is reserved before slicing.
+    Truncates on whichever limit is hit first (line count or UTF-8 byte size).
+    ``max_bytes`` bounds the entire joined result, notice and separators
+    included.
     """
     lines = text.split("\n")
     total_bytes = _byte_len(text)
     if len(lines) <= max_lines and total_bytes <= max_bytes:
-        return None
+        return text
 
-    # Reserve using the largest counts/path; ``+ 4`` covers the two "\n\n".
-    notice_overhead = (
-        max(
-            _byte_len(
-                template.format(
-                    lines=len(lines),
-                    bytes=total_bytes,
-                    path=_SAMPLE_WORKSPACE_PATH,
-                )
-            )
-            for template in notice_templates
-        )
-        + 4
-    )
+    # Reserve notice + separator bytes up front; ``+ 4`` covers the two "\n\n".
+    notice_overhead = _byte_len(_TRUNCATION_NOTICE.format(lines=len(lines), bytes=total_bytes)) + 4
     byte_budget = max(2, max_bytes - notice_overhead)
 
     head_lines = max(1, max_lines // 2)
@@ -121,46 +70,5 @@ def _head_tail(
     kept_lines = len(head.split("\n")) + (len(tail.split("\n")) if tail else 0)
     dropped_lines = max(0, len(lines) - kept_lines)
     dropped_bytes = max(0, total_bytes - _byte_len(head) - _byte_len(tail))
-    return head, tail, dropped_lines, dropped_bytes
-
-
-def _join(head: str, tail: str, notice: str) -> str:
+    notice = _TRUNCATION_NOTICE.format(lines=dropped_lines, bytes=dropped_bytes)
     return f"{head}\n\n{notice}\n\n{tail}" if tail else f"{head}\n\n{notice}"
-
-
-def bound_text(text: str, *, max_lines: int, max_bytes: int) -> str:
-    """Return ``text`` unchanged when small, else a head+tail preview.
-
-    Nothing is persisted; use :func:`bound_and_store` to keep the full output.
-    """
-    parts = _head_tail(text, max_lines, max_bytes)
-    if parts is None:
-        return text
-    head, tail, dropped_lines, dropped_bytes = parts
-    return _join(head, tail, _TRUNCATION_NOTICE.format(lines=dropped_lines, bytes=dropped_bytes))
-
-
-async def bound_and_store(text: str, *, max_lines: int, max_bytes: int) -> str:
-    """Like :func:`bound_text`, but spill the full output into the sandbox and
-    point the agent at its path. Degrades to a plain preview if the spill fails.
-    """
-    parts = _head_tail(
-        text,
-        max_lines,
-        max_bytes,
-        notice_templates=(_WORKSPACE_SPILL_NOTICE, _TRUNCATION_NOTICE),
-    )
-    if parts is None:
-        return text
-    head, tail, dropped_lines, dropped_bytes = parts
-
-    writer = _spill.get("writer")
-    if writer is not None:
-        path = await writer(uuid.uuid4().hex, text)
-        if path is not None:
-            notice = _WORKSPACE_SPILL_NOTICE.format(
-                lines=dropped_lines, bytes=dropped_bytes, path=path
-            )
-            return _join(head, tail, notice)
-
-    return _join(head, tail, _TRUNCATION_NOTICE.format(lines=dropped_lines, bytes=dropped_bytes))

strix/tools/reporting/tool.py

@@ -1,13 +1,7 @@
-"""Reporting tools — file vuln findings (with dedup + CVSS) and read them back.
-
-``create_vulnerability_report`` / ``create_dependency_report`` file findings;
-``list_reports`` / ``get_report`` let any agent (notably the root orchestrator)
-review what's been filed so far across the whole scan.
-"""
+"""``create_vulnerability_report`` — file a vuln finding with dedup + CVSS."""
 
 from __future__ import annotations
 
-import asyncio
 import json
 import logging
 import re
@@ -324,21 +318,6 @@ async def _do_create(  # noqa: PLR0912
         }
 
 
-def _caller_identity(ctx: RunContextWrapper) -> tuple[str | None, str | None]:
-    """Return the (agent_id, agent_name) of the agent invoking this tool."""
-    inner = ctx.context if isinstance(ctx.context, dict) else {}
-    raw_agent_id = inner.get("agent_id")
-    agent_id = raw_agent_id if isinstance(raw_agent_id, str) else None
-    agent_name: str | None = None
-    coordinator = inner.get("coordinator")
-    if agent_id is not None and coordinator is not None:
-        names = getattr(coordinator, "names", {})
-        if isinstance(names, dict):
-            raw_agent_name = names.get(agent_id)
-            agent_name = raw_agent_name if isinstance(raw_agent_name, str) else None
-    return agent_id, agent_name
-
-
 @function_tool(timeout=180, strict_mode=False)
 async def create_vulnerability_report(
     ctx: RunContextWrapper,
@@ -625,7 +604,16 @@ async def create_vulnerability_report(
             template engine's auto-escaping over string interpolation.
         fix_effort: "low"
     """
-    agent_id, agent_name = _caller_identity(ctx)
+    inner = ctx.context if isinstance(ctx.context, dict) else {}
+    raw_agent_id = inner.get("agent_id")
+    agent_id = raw_agent_id if isinstance(raw_agent_id, str) else None
+    agent_name = None
+    coordinator = inner.get("coordinator")
+    if agent_id is not None and coordinator is not None:
+        names = getattr(coordinator, "names", {})
+        if isinstance(names, dict):
+            raw_agent_name = names.get(agent_id)
+            agent_name = raw_agent_name if isinstance(raw_agent_name, str) else None
 
     result = await _do_create(
         title=title,
@@ -930,7 +918,16 @@ async def create_dependency_report(
         fix_effort: One of ``trivial`` / ``low`` / ``medium`` / ``high``
             (dependency upgrades are usually ``trivial``/``low``).
     """
-    agent_id, agent_name = _caller_identity(ctx)
+    inner = ctx.context if isinstance(ctx.context, dict) else {}
+    raw_agent_id = inner.get("agent_id")
+    agent_id = raw_agent_id if isinstance(raw_agent_id, str) else None
+    agent_name = None
+    coordinator = inner.get("coordinator")
+    if agent_id is not None and coordinator is not None:
+        names = getattr(coordinator, "names", {})
+        if isinstance(names, dict):
+            raw_agent_name = names.get(agent_id)
+            agent_name = raw_agent_name if isinstance(raw_agent_name, str) else None
 
     result = await _do_create_dependency(
         title=title,
@@ -952,288 +949,3 @@ async def create_dependency_report(
         agent_name=agent_name,
     )
     return json.dumps(result, ensure_ascii=False, default=str)
-
-
-_SEVERITY_ORDER = {
-    "critical": 0,
-    "high": 1,
-    "medium": 2,
-    "low": 3,
-    "info": 4,
-    "none": 5,
-}
-_VALID_SEVERITIES = frozenset(_SEVERITY_ORDER)
-_VALID_FINDING_CLASSES = frozenset({"dynamic", "dependency_cve"})
-_REPORT_DESCRIPTION_PREVIEW_CHARS = 280
-
-# Compact, listing-safe fields — no full bodies / PoC code / evidence.
-_REPORT_SUMMARY_FIELDS = (
-    "id",
-    "title",
-    "severity",
-    "cvss",
-    "finding_class",
-    "cve",
-    "cwe",
-    "target",
-    "endpoint",
-    "method",
-    "fix_effort",
-    "agent_name",
-    "timestamp",
-)
-
-
-def _report_severity_rank(report: dict[str, Any]) -> int:
-    return _SEVERITY_ORDER.get(str(report.get("severity", "")).lower(), 99)
-
-
-def _report_matches_filters(
-    report: dict[str, Any],
-    *,
-    severity: str | None,
-    finding_class: str | None,
-    target: str | None,
-    search: str | None,
-) -> bool:
-    if severity and str(report.get("severity", "")).lower() != severity:
-        return False
-    if finding_class and str(report.get("finding_class", "dynamic")).lower() != finding_class:
-        return False
-    if target:
-        target_lower = target.lower()
-        haystack = f"{report.get('target', '')} {report.get('endpoint', '')}".lower()
-        if target_lower not in haystack:
-            return False
-    if search:
-        search_lower = search.lower()
-        title_match = search_lower in str(report.get("title", "")).lower()
-        desc_match = search_lower in str(report.get("description", "")).lower()
-        if not (title_match or desc_match):
-            return False
-    return True
-
-
-def _mark_authorship(
-    entry: dict[str, Any], report: dict[str, Any], caller_agent_id: str | None
-) -> dict[str, Any]:
-    """Flag whether ``report`` was filed by the agent making this call."""
-    if caller_agent_id is not None and report.get("agent_id") == caller_agent_id:
-        entry["by_you"] = True
-    return entry
-
-
-def _to_report_summary_entry(
-    report: dict[str, Any], caller_agent_id: str | None = None
-) -> dict[str, Any]:
-    entry = {
-        field: report[field] for field in _REPORT_SUMMARY_FIELDS if report.get(field) is not None
-    }
-    description = str(report.get("description", "")).strip()
-    if description:
-        if len(description) > _REPORT_DESCRIPTION_PREVIEW_CHARS:
-            entry["description_preview"] = (
-                f"{description[:_REPORT_DESCRIPTION_PREVIEW_CHARS].rstrip()}..."
-            )
-        else:
-            entry["description_preview"] = description
-    return _mark_authorship(entry, report, caller_agent_id)
-
-
-def _severity_counts(reports: list[dict[str, Any]]) -> dict[str, int]:
-    counts: dict[str, int] = {}
-    for report in reports:
-        sev = str(report.get("severity", "")).lower() or "none"
-        counts[sev] = counts.get(sev, 0) + 1
-    return {sev: counts[sev] for sev in _SEVERITY_ORDER if sev in counts}
-
-
-async def _run_report_reader(fn: Any, *args: Any, **kwargs: Any) -> dict[str, Any]:
-    try:
-        return await asyncio.to_thread(fn, *args, **kwargs)
-    except (ImportError, AttributeError) as e:
-        logger.exception("report reader failed")
-        return {"success": False, "error": f"Failed to read reports: {e!s}"}
-
-
-def _do_list_reports(
-    *,
-    severity: str | None,
-    finding_class: str | None,
-    target: str | None,
-    search: str | None,
-    include_details: bool,
-    caller_agent_id: str | None = None,
-) -> dict[str, Any]:
-    errors: list[str] = []
-    severity = (severity or "").strip().lower() or None
-    if severity and severity not in _VALID_SEVERITIES:
-        errors.append(
-            f"Invalid severity: {severity!r}. Must be one of: {sorted(_VALID_SEVERITIES)}"
-        )
-    finding_class = (finding_class or "").strip().lower() or None
-    if finding_class and finding_class not in _VALID_FINDING_CLASSES:
-        errors.append(
-            f"Invalid finding_class: {finding_class!r}. "
-            f"Must be one of: {sorted(_VALID_FINDING_CLASSES)}"
-        )
-    if errors:
-        return {"success": False, "error": "Validation failed", "errors": errors}
-
-    from strix.report.state import get_global_report_state
-
-    report_state = get_global_report_state()
-    if report_state is None:
-        return {
-            "success": True,
-            "reports": [],
-            "filtered_count": 0,
-            "total_count": 0,
-            "severity_counts": {},
-            "warning": "Report state unavailable - no reports have been filed yet",
-        }
-
-    all_reports = report_state.get_existing_vulnerabilities()
-    matched = [
-        r
-        for r in all_reports
-        if _report_matches_filters(
-            r,
-            severity=severity,
-            finding_class=finding_class,
-            target=(target or "").strip() or None,
-            search=(search or "").strip() or None,
-        )
-    ]
-    matched.sort(key=lambda r: (_report_severity_rank(r), str(r.get("id", ""))))
-
-    reports = [
-        _mark_authorship(dict(r), r, caller_agent_id)
-        if include_details
-        else _to_report_summary_entry(r, caller_agent_id)
-        for r in matched
-    ]
-    return {
-        "success": True,
-        "reports": reports,
-        "filtered_count": len(reports),
-        "total_count": len(all_reports),
-        "severity_counts": _severity_counts(all_reports),
-    }
-
-
-def _do_get_report(report_id: str, caller_agent_id: str | None = None) -> dict[str, Any]:
-    report_id = (report_id or "").strip()
-    if not report_id:
-        return {"success": False, "error": "report_id cannot be empty", "report": None}
-
-    from strix.report.state import get_global_report_state
-
-    report_state = get_global_report_state()
-    if report_state is None:
-        return {
-            "success": False,
-            "error": "Report state unavailable - no reports have been filed yet",
-            "report": None,
-        }
-
-    for report in report_state.get_existing_vulnerabilities():
-        if report.get("id") == report_id:
-            return {
-                "success": True,
-                "report": _mark_authorship(dict(report), report, caller_agent_id),
-            }
-    return {
-        "success": False,
-        "error": f"Report with id '{report_id}' not found",
-        "report": None,
-    }
-
-
-@function_tool(timeout=30)
-async def list_reports(
-    ctx: RunContextWrapper,
-    severity: str | None = None,
-    finding_class: str | None = None,
-    target: str | None = None,
-    search: str | None = None,
-    include_details: bool = False,
-) -> str:
-    """List vulnerability reports filed so far in this scan — metadata-first.
-
-    **For the orchestrator / root agent.** This is an orchestration tool
-    for tracking scan-wide coverage and assembling the final report — leaf
-    / specialist agents do their own testing and file findings; they should
-    NOT call this. If you are a subagent, ignore it and focus on your task.
-
-    Reports are shared across **every** agent in the scan, so this returns
-    findings filed by any agent (root or child), not just your own. As the
-    root agent, use it to track progress, avoid dispatching work on
-    already-covered ground, reason about attack-chaining across confirmed
-    findings, and build the ``finish_scan`` executive summary.
-
-    By default each entry is compact: ``id``, ``title``, ``severity``,
-    ``cvss``, ``finding_class``, ``cve`` / ``cwe``, ``target`` /
-    ``endpoint``, ``fix_effort``, ``agent_name`` (who filed it), ``timestamp``,
-    plus a 280-char ``description_preview``. Entries you filed yourself are
-    flagged ``by_you: true``. The response also carries
-    ``total_count`` and ``severity_counts`` (counts per severity across all
-    reports, ignoring filters). Set ``include_details=True`` for full report
-    bodies (PoC, evidence, remediation, code_locations) — token-expensive;
-    prefer ``get_report`` to drill into a single finding.
-
-    Filters compose (all must match): ``severity`` and ``finding_class``
-    match exactly, ``target`` is a substring match against target/endpoint,
-    and ``search`` is a substring match against title/description. Results
-    are ordered by severity (critical -> info), then report id.
-
-    This is read-only — it never files or dedupes anything.
-
-    Args:
-        severity: Filter to one of ``critical`` / ``high`` / ``medium`` /
-            ``low`` / ``info`` / ``none``.
-        finding_class: Filter to ``dynamic`` (PoC-backed) or
-            ``dependency_cve`` (known-CVE supply-chain).
-        target: Substring match against a report's target / endpoint.
-        search: Substring match against title and description.
-        include_details: When False (default) entries are compact; when
-            True full report bodies are returned.
-    """
-    caller_agent_id, _ = _caller_identity(ctx)
-    return json.dumps(
-        await _run_report_reader(
-            _do_list_reports,
-            severity=severity,
-            finding_class=finding_class,
-            target=target,
-            search=search,
-            include_details=include_details,
-            caller_agent_id=caller_agent_id,
-        ),
-        ensure_ascii=False,
-        default=str,
-    )
-
-
-@function_tool(timeout=30)
-async def get_report(ctx: RunContextWrapper, report_id: str) -> str:
-    """Fetch one vulnerability report by its id (e.g. ``vuln-0001``).
-
-    Returns the full report body — description, impact, technical analysis,
-    PoC, evidence, remediation, CVSS breakdown, and any ``code_locations``.
-    Use ``list_reports`` first to find ids; this is the cheap way to read a
-    single finding in full without pulling every body.
-
-    Read-only.
-
-    Args:
-        report_id: Report id from ``list_reports`` or a
-            ``create_vulnerability_report`` / ``create_dependency_report``
-            response (format ``vuln-NNNN``).
-    """
-    caller_agent_id, _ = _caller_identity(ctx)
-    return json.dumps(
-        await _run_report_reader(_do_get_report, report_id, caller_agent_id),
-        ensure_ascii=False,
-        default=str,
-    )

strix/viewer/__init__.py

@@ -6,7 +6,7 @@ directly from the run's on-disk files. No cloud dependency, no file picker.
 
 from __future__ import annotations
 
-from strix.interface.viewer.server import serve
+from strix.viewer.server import serve
 
 
 __all__ = ["serve"]

strix/viewer/auth.py

@@ -155,7 +155,7 @@ def _post_json(path: str, payload: dict[str, Any], *, timeout: int) -> tuple[int
         method="POST",
     )
     try:
-        with urllib.request.urlopen(request, timeout=timeout) as response:  # noqa: S310  # nosec B310
+        with urllib.request.urlopen(request, timeout=timeout) as response:  # noqa: S310
             return response.status, _parse_body(response.read())
     except urllib.error.HTTPError as exc:
         return exc.code, _parse_body(exc.read())

strix/viewer/cli.py

@@ -16,8 +16,8 @@ from strix.core.paths import (
     run_record_path,
     runs_base_dir,
 )
-from strix.interface.viewer.server import authorized_url, bundle_is_built, serve
-from strix.interface.viewer.transcript import read_run_summary
+from strix.viewer.server import authorized_url, bundle_is_built, serve
+from strix.viewer.transcript import read_run_summary
 
 
 if TYPE_CHECKING:
@@ -58,7 +58,7 @@ def run_view(argv: list[str]) -> None:
     if not bundle_is_built():
         console.print(
             "[bold red]Viewer UI is not built.[/]\n"
-            "Build it with: [cyan]cd strix/interface/viewer/frontend && npm ci && npm run build[/]"
+            "Build it with: [cyan]cd strix/viewer/frontend && npm ci && npm run build[/]"
         )
         raise SystemExit(1)
 

strix/viewer/frontend/src/components/live/tool-renderers/NotesRenderer.tsx

@@ -49,9 +49,6 @@ export default function NotesRenderer({ toolName, args, result }: ToolRendererPr
             <div className="mt-1.5 text-[#999] text-[13px]">
               {note.title ?? "(untitled)"}
               <span className="text-[#555] ml-1">({note.category ?? "general"})</span>
-              {(note.by_you || note.agent_name) && (
-                <span className="text-[#666] ml-1 text-xs">by {note.by_you ? "you" : note.agent_name}</span>
-              )}
             </div>
             {note.content && <div className="mt-1"><Markdown text={note.content} /></div>}
           </>
@@ -77,9 +74,6 @@ export default function NotesRenderer({ toolName, args, result }: ToolRendererPr
                 <span className="text-[#555] mr-1">-</span>
                 <span className="text-[#999]">{n.title ?? "(untitled)"}</span>
                 <span className="text-[#555] ml-1">({n.category ?? "general"})</span>
-                {(n.by_you || n.agent_name) && (
-                  <span className="text-[#666] ml-1 text-xs">by {n.by_you ? "you" : n.agent_name}</span>
-                )}
                 {n.content && <div className="ml-3"><Markdown text={n.content} /></div>}
               </div>
             ))}

strix/viewer/frontend/src/components/live/tool-renderers/index.ts

@@ -12,7 +12,6 @@ import FileEditRenderer from "./FileEditRenderer";
 import ApplyPatchRenderer from "./ApplyPatchRenderer";
 import ViewImageRenderer from "./ViewImageRenderer";
 import VulnReportRenderer from "./VulnReportRenderer";
-import ReportListRenderer from "./ReportListRenderer";
 import ProxyRenderer from "./ProxyRenderer";
 import ThinkRenderer from "./ThinkRenderer";
 import AgentCommsRenderer from "./AgentCommsRenderer";
@@ -102,7 +101,7 @@ const CATEGORY_TOOLS: Record<ToolCategory, readonly string[]> = {
   filesystem: ["apply_patch", "view_image", "str_replace_editor", "list_files", "search_files"],
   // Caido proxy tools (legacy: send_request)
   proxy: ["list_requests", "view_request", "repeat_request", "list_sitemap", "view_sitemap_entry", "scope_rules", "send_request"],
-  reporting: ["create_vulnerability_report", "list_reports", "get_report"],
+  reporting: ["create_vulnerability_report"],
   thinking: ["think"],
   agents: ["create_agent", "agent_finish", "send_message_to_agent", "wait_for_message", "view_agent_graph", "stop_agent"],
   search: ["web_search"],
@@ -129,8 +128,6 @@ const RENDERER_OVERRIDES: Partial<Record<string, ComponentType<ToolRendererProps
   finish_scan: FinishRenderer,
   apply_patch: ApplyPatchRenderer,
   view_image: ViewImageRenderer,
-  list_reports: ReportListRenderer,
-  get_report: ReportListRenderer,
 };
 
 /**