Repository navigation
chore: remove Entire trail runners - #1485
Conversation
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
|
Navigate logical layers of code changes, visualize relationships, and explore their blast radius. No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info⚙️ Run configurationConfiguration used: Central YAML (base), Organization UI (inherited) Review profile: CHILL Plan: Advanced Run ID: 📒 Files selected for processing (6)
💤 Files with no reviewable changes (6)
Included review availability: Your plan provides up to 4 included reviews per hour; 2 remain after this review. 📝 WalkthroughWalkthroughSix runner configuration files were deleted. They defined trail review, summary, confidence, drift, risk, and security automations. ChangesTrail runner configurations
Priority: ⬇️ Low Estimated code review effort: 2 (Simple) | ~10 minutes Change: Other Merge Risk: ⚪ Minimal · up to The Entire automations will stop running as intended. No additional merge-blocking concern is established. 🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 passed)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
|
Instruction countsThe comparison never ran — an earlier step failed.
|
<!-- entire-trail-link-start --> https://entire.io/gh/jdx/usage/trails/33 <!-- entire-trail-link-end --> ### 🚀 Features - **(cli)** publish a usage agent skill with packslip by [@jdx](https://github.com/jdx) in [#1509](#1509) - **(complete)** complete a wrapped command's arguments with its own shell completion via delegate= by [@jdx](https://github.com/jdx) in [#1498](#1498) - **(lib)** add getters that read FlagMeta, CommandMeta and ArgMeta fields wherever they live by [@jdx](https://github.com/jdx) in [#1491](#1491) - **(lib)** export the chosen subcommand to scripts as usage_cmd by [@jdx](https://github.com/jdx) in [#1505](#1505) - **(spec)** let a complete node sit inside the arg it completes by [@jdx](https://github.com/jdx) in [#1496](#1496) - **(spec)** list an arg's possible values from a command with `choices run=` by [@jdx](https://github.com/jdx) in [#1497](#1497) ### 🐛 Bug Fixes - **(bash)** complete `--flag=` values when the cursor sits after `=` by [@jdx](https://github.com/jdx) in [#1499](#1499) - **(cli)** run scripts passed to usage bash through a pipe or process substitution by [@jdx](https://github.com/jdx) in [#1494](#1494) - **(cli)** keep usage explain from running a spec's choices run= commands by [@jdx](https://github.com/jdx) in [#1503](#1503) - **(complete)** read --line=LINE in completion requests from typed completers by [@jdx](https://github.com/jdx) in [#1487](#1487) - **(complete)** show completion parse errors without garbling the prompt by [@jdx](https://github.com/jdx) in [#1495](#1495) - **(derive)** ignore closed pipes instead of panicking in generated parse() by [@jdx](https://github.com/jdx) in [#1484](#1484) - **(derive)** keep generated async dispatch from reserving stack for every command by [@jdx](https://github.com/jdx) in [#1488](#1488) - **(fish)** complete words the user has started quoting by [@jdx](https://github.com/jdx) in [#1500](#1500) - **(lib)** resolve inherited usage aliases after multiline workspace entries by [@jdx](https://github.com/jdx) in [#1507](#1507) - **(lib)** make published usage-rs tests self-contained by [@jdx](https://github.com/jdx) in [#1510](#1510) - **(zsh)** stop the completion-init handler from breaking file completion for other commands by [@jdx](https://github.com/jdx) in [#1493](#1493) ### 📚 Documentation - **(cli)** describe fig output as the legacy Fig format by [@jdx](https://github.com/jdx) in [2a1bef6](2a1bef6) - clarify CLI frameworks and generators on the homepage by [@jdx](https://github.com/jdx) in [#1482](#1482) ### 🛡️ Security - remove Entire trail runners by [@jdx](https://github.com/jdx) in [#1485](#1485) ### 🔍 Other Changes - float jdx tools and aube on latest without a release-age delay by [@jdx](https://github.com/jdx) in [#1486](#1486) - run cargo-semver-checks on every published crate by [@jdx](https://github.com/jdx) in [#1490](#1490) - fail pull requests that grow the usage CLI or a derived CLI by more than 1% by [@jdx](https://github.com/jdx) in [#1492](#1492) - make the binary-size check measure the pull request's own code by [@jdx](https://github.com/jdx) in [#1501](#1501) ### 📦️ Dependency Updates - bump jdx/renovate-config workflows to c736149 by [@jdx](https://github.com/jdx) in [06f2f1b](06f2f1b) - bump jdx/renovate-config workflows to aa49efc by [@jdx](https://github.com/jdx) in [9aaedef](9aaedef) - bump jdx/renovate-config workflows to 5b46432 by [@jdx](https://github.com/jdx) in [36f8681](36f8681) - pin jdx/renovate-config workflows to v1.0.0 by [@jdx](https://github.com/jdx) in [a31ae2f](a31ae2f) - update communique to 1.4.2 in mise.lock by [@jdx](https://github.com/jdx) in [805396f](805396f) - upgrade locked mise tools by [@jdx](https://github.com/jdx) in [d8b761d](d8b761d) - update jdx/packslip action to v1.4.0 by [@jdx](https://github.com/jdx) in [#1508](#1508)
https://entire.io/gh/jdx/usage/trails/35
Removes the Entire trail runners under
.entire/runners/(trail-review,trail-summary,trail-confidence,trail-risk,trail-security,trail-drift). Branches will no longer get automated Entire review comments, summaries, or confidence/risk/security/drift scores, and the "Entire Gates" check stops running.Entire itself stays enabled:
.entire/settings.json(session checkpoints) is unchanged. Configuration only; no code or runtime behavior changes.🤖 Generated with Claude Code
Note
Low Risk
Deletes CI/automation JSON only; no library or runtime code paths change, though teams lose automated Entire trail checks until runners are restored elsewhere.
Overview
Removes all Entire trail-scoped prompt runners from
.entire/runners/: automated branch review (trail-review), markdown trail summary (trail-summary), and monitor evals for confidence, risk, security, and drift.Those configs defined Claude prompt runners (timeouts, sandbox, triggers on push/API) and how results were adapted (code review comments, trail monitors, markdown). After this change, branches no longer get Entire-driven review comments, summaries, or those score monitors, and the associated “Entire Gates” automation stops.
Entire session checkpoints stay on via unchanged
.entire/settings.json; this is configuration cleanup only, with no application code changes.Reviewed by Cursor Bugbot for commit 681b413. Bugbot is set up for automated code reviews on this repo. Configure here.
Summary by CodeRabbit