Repository navigation
Add autonomously-fix-small-issue skill - #935
Merged
Merged
Conversation
Codifies the workflow used for GH #930 / PR #934: read the issue, pass a scope gate, file a braid strand and plan doc, branch, fix test-first, verify (separating pre-existing failures from new ones), and open a PR. The name flags that the run is pre-approved and unattended. The safeguards are structural instead: user-invoked only (disable-model-invocation), a strict scope gate that hands off to /triage, and a hard stop at the PR (never merge or close the issue or strand). Strand: bd-t18wrsqo. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Contributor
✅ Snyk checks have passed. No issues have been found so far.
💻 Catch issues earlier using the plugins for VS Code, JetBrains IDEs, Visual Studio, and Eclipse. |
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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
/autonomously-fix-small-issue <N>, which codifies the workflow from #930 → #934. It takes a small, unambiguous GitHub issue all the way to an open PR without asking the user anything along the way.The name makes the risk obvious: once invoked, the run is unattended. The safeguards are built into the skill rather than relying on step-by-step confirmation:
disable-model-invocation: truemeans the agent never picks it on its own./triage.main) from new ones, and says what wasn't verified.main. Otherwise it usescargo xtask create-worktree. It never runsswitch-taskin the main checkout, because that rewritesCLAUDE.local.md.Testing: this PR adds documentation only, with no code changes. I didn't run
cargo xtask verify. The skill hasn't been exercised as a skill yet; the reference run (#934) was done by hand.Strand: bd-t18wrsqo · Plan:
claude-notes/plans/2026-10-27-autonomously-fix-small-issue-skill.md🤖 Generated with Claude Code