Adds the missing execution-validator test for the episode-identity guard.
Completion by reusing a valid same-task observation already succeeds when the
lease and the grounding trace share an episode (see
test_executor_can_complete_by_exactly_reusing_same_task_evidence). This adds
the mirror negative case: the same exact-quote DONE delta and reusable
ObservationRecord, but with the lease bound to a different trace_episode_id
than the trace carries. compile_execution refuses it with
evidence trace belongs to a different episode before any grounding is
credited — so evidence captured in another episode cannot float a completion.
Covers the execution.py:301-305 guard in the DONE-would-otherwise-succeed
path, which test_evidence_trace_must_match_the_persisted_episode_identity
(PROGRESS, empty events) did not exercise.
Verification:
uv run python -m pytest -q→ 120 passed, 1 skippeduv run ruff check src tests/ruff format --checkcleanuv run mypy srcclean
🤖 Generated with Claude Code
Activity
4
hacktivistxt9d agoConfirmed green: ran the same CI workflow on my fork for this exact SHA (ed1d6dd) — success: https://github.com/hacktivistxt/PentestGPT/actions/runs/41072974821 (pytest 120 passed/1 skipped, ruff check + format, mypy, all clean). This PR's run is just awaiting maintainer approval of the fork-PR workflow gate.
CI should pass on this branch. Locally against
pentestgpt_agent:uv run python -m pytest -q→ 120 passed, 1 skipped;uv run ruff check src testsanduv run ruff format --check src testsclean;uv run mypy srcclean. The change is test-only (adds one execution-validator case), so no runtime behavior is touched.