feat: autonomous weekly strategy agent (Claude Max, CI-gated)#201
Merged
Conversation
scripts/autonomous_review.py: the weekly agent proposes at most one evidence-backed change (read-only Claude on the Max subscription). The change is applied and shipped ONLY through the existing guarded deploy_change pipeline (budget, sample-size, safety invariants, pytest, 365-day backtest Sharpe>=0.3), then squash-merged after the required CI checks pass. - The LLM only proposes (read-only JSON); all file mutation and the git/merge steps are deterministic and routed through existing guardrails. - scripts/weekly_autonomous_agent.sh: cron wrapper, Max OAuth, passes --dry-run/--selftest. - safety.py: add the agent's own machinery (orchestrator, wrapper, runners) to the no-touch list, with matching test_safety assertions. Modes: --dry-run (propose->gates->report->revert, no git); --selftest (verify the no-touch wall rejects bad edits).
…kill switch, main branches) Mocks the subprocess boundaries (Claude, git, gh, deploy) so no real calls fire. Brings scripts/autonomous_review.py to 87% (>= 85% per-file floor).
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Adds the fully-autonomous weekly agent. The LLM (read-only, Max subscription) proposes at most one evidence-backed change; it is applied and shipped only through the existing guarded
deploy_changepipeline (budget, sample-size, safety invariants, pytest, 365-day backtest), then squash-merged after the required CI checks pass.Modes:
--dry-run(propose -> gates -> report -> revert, no git),--selftest(verify the no-touch wall rejects bad edits). The agent's own machinery is added to the no-touch list.NOT armed yet: the launchd weekly job is still the read-only reviewer; flipping it to this is a follow-up after dry-run review.