improve(create-pr): optimize token efficiency and fix bugs via autoresearch#608
improve(create-pr): optimize token efficiency and fix bugs via autoresearch#608
Conversation
…dundant comments
Result: {"status":"keep","total_bytes":6268,"line_count":183,"word_count":917}
…tighten prose
Result: {"status":"keep","total_bytes":5836,"line_count":169,"word_count":844}
Not referenced by any skill or hook. Was retained for pr-pass per spec
but pr-pass doesn't actually use it.
Result: {"status":"keep","total_bytes":4534,"line_count":141,"word_count":666}
…ync when behind)
Result: {"status":"keep","total_bytes":4019,"line_count":122,"word_count":601}
Only preflight-check.sh used lib.sh. Inlining saves a file and overhead.
Result: {"status":"keep","total_bytes":3558,"line_count":107,"word_count":541}
… inline notes
Result: {"status":"keep","total_bytes":3202,"line_count":85,"word_count":490}
Result: {"status":"keep","total_bytes":3103,"line_count":82,"word_count":471}
Result: {"status":"keep","total_bytes":2884,"line_count":68,"word_count":448}
Scripts don't load into LLM context — only SKILL.md token cost matters.
Result: {"status":"keep","skill_bytes":1081,"skill_lines":27,"skill_words":151}
…edundant sections
Remove stop conditions (obvious), trim CI failure instructions,
shorten description field.
Result: {"status":"keep","skill_bytes":802,"skill_lines":23,"skill_words":110}
Result: {"status":"keep","skill_bytes":732,"skill_lines":19,"skill_words":100}
Result: {"status":"keep","skill_bytes":675,"skill_lines":18,"skill_words":93}
Result: {"status":"keep","skill_bytes":635,"skill_lines":18,"skill_words":82}
Result: {"status":"keep","skill_bytes":605,"skill_lines":17,"skill_words":82}
…r-optimize-20260402
… preflight
Test found: auto-merge is disabled after force-push, must re-enable.
Also: preflight sync push needs -u for new branches.
Result: {"status":"keep","skill_bytes":665,"skill_lines":17,"skill_words":92}
Test showed agent skipping preflight-check.sh and doing manual logic instead.
Added "Execute each line literally" + "scripts MUST be run" directive.
Result: {"status":"keep","skill_bytes":794,"skill_lines":20,"skill_words":109}
- Remove tests for deleted scripts (lib.sh, verify-pr-status.sh, sync-with-base.sh)
- Update preflight BEHIND test to expect auto-sync behavior
- Fix skill content test to accept code-block-only skills
Result: {"status":"keep","skill_bytes":794}
…erge
Aligns with SKILL.md's gh pr merge --auto --squash (no delete-branch).
Prevents unexpected branch deletion on fallback merge path.
Result: {"status":"keep","skill_bytes":794,"script_bytes":1802}
…heck
Branch created from main is always up-to-date with base — re-run is unnecessary.
Move main/master check before preflight for logical ordering.
Result: {"status":"keep","skill_bytes":776,"skill_lines":20,"skill_words":107}
Result: {"status":"keep","skill_bytes":776}
…r-optimize-20260402 # Conflicts: # .autoresearch/autoresearch.jsonl # .autoresearch/autoresearch.md # .autoresearch/dashboard.md # .autoresearch/worklog.md # plugins/me/skills/create-pr/SKILL.md # plugins/me/skills/create-pr/scripts/wait-for-merge.sh
|
Warning Rate limit exceeded
Your organization is not enrolled in usage-based pricing. Contact your admin to enable usage-based pricing to continue reviews beyond the rate limit, or try again in 16 minutes and 58 seconds. ⌛ How to resolve this issue?After the wait time has elapsed, a review can be triggered using the We recommend that you space out your commits to avoid hitting the rate limit. 🚦 How do rate limits work?CodeRabbit enforces hourly rate limits for each developer per organization. Our paid plans have higher rate limits than the trial, open-source and free plans. In all cases, we re-allow further reviews after a brief timeout. Please see our FAQ for further information. ℹ️ Review info⚙️ Run configurationConfiguration used: defaults Review profile: CHILL Plan: Pro Run ID: 📒 Files selected for processing (11)
✨ Finishing Touches🧪 Generate unit tests (beta)
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 |
Summary
Optimize the
create-prskill for token efficiency (SKILL.md -59%) and fix 5 bugs found throughsubagent testing (4 real PRs created and merged as validation).
Changes
Structural (file reduction: 6→3 files)
verify-pr-status.sh(not referenced by any skill)sync-with-base.shintopreflight-check.sh(auto-sync when behind)lib.shintopreflight-check.sh(single consumer)SKILL.md optimization (1,898→776 bytes, -59%)
S=path variable for script pathsBug fixes (found via 4 subagent PR tests)
preflight-check.sh: usegit push -ufor new branches without upstreamwait-for-merge.sh: remove--delete-branchfrom fallback merge (inconsistent with SKILL.md)Tests
bats tests/— 63/63 passpre-commit run --all-files— passshellcheck -x— all scripts pass