Validate vehicle-level hazard mappings and reasoning for each Function_ID.
npx skills add https://github.com/WSJ40/hara-mutiagent --skill hara-stage2r
Please help me install this Agent Skill. Skill: hara-stage2r Source: https://github.com/WSJ40/hara-mutiagent/tree/main/skills/hara-stage2r Command: npx skills add https://github.com/WSJ40/hara-mutiagent --skill hara-stage2r
This Skill helps you perform a human-grade semantic review of vehicle-level hazard mappings and reasoning for each Function_ID after Stage 2 single-function machine checks pass, ensuring the final Stage2 merge is only allowed when hazard semantics and fault descriptions are usable for downstream Stage 3.
Run the Stage2 machine check for the target Function_ID, then manually review the hazards and reasoning using the rules in references/stage2-review.md, and finally write the Function_ID-specific review JSON to determine whether the slice is allowed into the final merge.
High-intent search queries and answers about installing and using this skill.
To perform a semantic hazard review, run Stage 2 machine checks for the target Function_ID first, then manually verify hazard mappings, hazard_reasoning, and trace fields using stage2-review rules to ensure fault descriptions are suitable for Stage 3.
A Function_ID-identifiable JSON review file records pass or fail decisions with specific issues and fixes, allowing merge_stage2_review.py to aggregate the review traces and enforce the Stage 2 merge gate.
You need to validate vehicle-level hazard semantics after Stage 2 machine checks pass, specifically when reviewing hazard mapping credibility and fault-description boundaries, while rejecting schema or mechanical issues to stage2_slice --fix.
You cannot proceed to the Stage 2 merge if hazard reasoning and traceability checks fail. The merge gate enforcement requires all Function_ID Stage2R semantic reviews to pass before allowing the final Stage 2 merge.
Traceability validation during Stage 2 gating works by re-checking hazard mapping selection and hazard_reasoning credibility for a single Function_ID slice, ensuring that trace fields and fault-description boundaries are usable for downstream Stage 3 analysis.
The limitation of applying stage2-review is that it only handles semantic validation for a single Function_ID slice and cannot resolve mechanical schema issues, which must be routed to stage2_slice --fix before the semantic review begins.
djankiesGuide high-risk changes through plan, validate, execute, and verify phases.
WSJ40Convert Stage 1 safety faults into vehicle-level hazard JSON entries per Function_ID.
ssdavidaiRepair vault records by fixing frontmatter, orphaned records, and content issues.
insightwavesmithAnalyze code changes to identify those meeting evidence-shape-check criteria.
