fix(safety): do not treat data tools reading script-named files as execution

Trace review of two guarded-mode runs found the unresolved-execution guard
firing on ordinary commands: curl fetching a .js asset, rg over a .py file, sed
and cat and cp on script-named files — 24 blocks in one run. The guard was meant
to fail closed on an unknown interpreter handed a script, but a read, transfer,
or text tool takes such a file as data, not as a program to run.

Exclude known read commands, HTTP clients, and a set of text/data tools from the
script-suffix branch, so only a genuinely unknown executable given a script still
fails closed. awk moves from the interpreter set to the data tools: its program
is an inline positional argument, not a -c or script file the entrypoint reader
can resolve.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>

tests/test_safety_evidence.py

@@ -898,3 +898,21 @@ async def test_oversize_input_file_is_attached_truncated() -> None:
         assert inp["bytes"] <= settings.max_artifact_bytes
     finally:
         bundle.cleanup()
+
+
+@pytest.mark.parametrize(
+    "command",
+    [
+        "curl -sS https://example.test/js/app.js -o /dev/null",
+        "wget https://example.test/main.bundle.js -O out.js",
+        "rg -n 'pattern' /workspace/app.py",
+        "sed -n '1,20p' /workspace/probe.py",
+        "cat /workspace/onboarding.py",
+        "cp /workspace/a.sh /workspace/b.sh",
+        "awk '{print $1}' /workspace/hosts.py",
+    ],
+)
+def test_data_tools_reading_script_named_files_are_not_execution(command: str) -> None:
+    """A read/transfer/text tool takes a script-named file as data, not as a program to
+    run, so it must not trip the unresolved-execution guard."""
+    assert parse_command(command).parse_error is None