Final polish: non-root Dockerfile, /reset task_id API support, logging cleanup, new tests#2

Merged

bigturtle679 merged 13 commits into

mainfrom

copilot/improve-inference-script-again

Apr 11, 2026

Commits on Apr 11, 2026

Add 3 new tasks, opponent simulation, enhanced grading, and comprehensive tests

Copilot
and
AbeerChaturvedi
authored
Address code review feedback: fix help text and comment notation

Copilot
and
AbeerChaturvedi
authored
Fix grader bugs, add opponent-aware negotiation, HTTP-client mode, and action_space

Copilot
and
AbeerChaturvedi
authored
Address code review: use task.trap_markers generically in _risk_score, clarify comments

Copilot
and
AbeerChaturvedi
authored
Add concession tracking, --retry-low flag, smarter ACCEPT gate, silence pydantic warnings

Copilot
and
AbeerChaturvedi
authored
Fix stale README test count (51), extract magic numbers to named constants

Copilot
and
AbeerChaturvedi
authored
Harden validation, fix 3 failing tests, add Literal types, edge-case tests, input length guards

Copilot
and
AbeerChaturvedi
authored
Consolidate _VALID_ACTION_TYPES to derive from canonical ActionType Literal in models.py

Copilot
and
AbeerChaturvedi
authored
Fix critical accept-blocking bug, retry logic, stack trace leak, session leak, pin deps

Copilot
and
AbeerChaturvedi
authored
Improve _is_negated docstring for clarity

Copilot
and
AbeerChaturvedi
authored
Replace assertions with proper validation, add Pydantic request model, docstrings, new API tests

Copilot
and
AbeerChaturvedi
authored
Harden error handling, add docstrings, HTTP timeouts, validation, and 3 new tests

Copilot
and
AbeerChaturvedi
authored
Final polish: non-root Dockerfile, /reset task_id support, logging fix, 3 new tests

Copilot
and
AbeerChaturvedi
authored

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Final polish: non-root Dockerfile, /reset task_id API support, logging cleanup, new tests#2

Final polish: non-root Dockerfile, /reset task_id API support, logging cleanup, new tests#2
bigturtle679 merged 13 commits into
mainfrom
copilot/improve-inference-script-again

Commits on Apr 11, 2026

Add 3 new tasks, opponent simulation, enhanced grading, and comprehensive tests

Address code review feedback: fix help text and comment notation

Fix grader bugs, add opponent-aware negotiation, HTTP-client mode, and action_space

Address code review: use task.trap_markers generically in _risk_score, clarify comments

Add concession tracking, --retry-low flag, smarter ACCEPT gate, silence pydantic warnings

Fix stale README test count (51), extract magic numbers to named constants

Harden validation, fix 3 failing tests, add Literal types, edge-case tests, input length guards

Consolidate _VALID_ACTION_TYPES to derive from canonical ActionType Literal in models.py

Fix critical accept-blocking bug, retry logic, stack trace leak, session leak, pin deps

Improve _is_negated docstring for clarity

Replace assertions with proper validation, add Pydantic request model, docstrings, new API tests

Harden error handling, add docstrings, HTTP timeouts, validation, and 3 new tests

Final polish: non-root Dockerfile, /reset task_id support, logging fix, 3 new tests