chore(skills): parallelize /precheck with model-routed sub-agents#572
Merged
Conversation
Restructures the /precheck procedure into explicit serial+parallel steps: - Step 1: ibm() stays inline (one MCP call, hard stop on fail) - Step 2: Jobs A–D fire in parallel as Agent calls in a single message - Job A: spec_validate (Haiku) — H2 section detection - Job B: repo validation (Haiku) — reads .claude-project.md for toolchain - Job C: trivy scan (Haiku) — run + parse JSON - Job D: code-reviewer (Opus) — confidence-calibrated quality gate - Steps 3–5: fix findings, assemble checklist, notify, gate (unchanged) Reduces wall-clock time from ~2-4 min serial to the longest single leg, and keeps tool results out of the main context window. Closes #571
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
Restructures the
/precheckskill procedure to run its four verification jobs in parallel sub-agents with explicit model selection, reducing wall-clock time and keeping tool results out of the main context window.Changes
spec_validate_structure— Haiku (mechanical H2 section detection).claude-project.mdfor toolchain first, then probes)Linked Issues
Closes #571
Test Plan
./scripts/ci/validate.sh— PASS (124 checks: shellcheck, shfmt, Python syntax, SKILL.md frontmatter, regression tests).claude-project.mdbefore blind-probing toolchain)