Derive Stage 1 function-level malfunction candidates from Stage 0 mappings.
npx skills add https://github.com/WSJ40/hara-mutiagent --skill hara-stage1
Please help me install this Agent Skill. Skill: hara-stage1 Source: https://github.com/WSJ40/hara-mutiagent/tree/main/skills/hara-stage1 Command: npx skills add https://github.com/WSJ40/hara-mutiagent --skill hara-stage1
This Skill helps you derive Stage 1 function-level malfunction entries from Stage 0 function mappings, ensuring every output row is traceable to each function’s detail_text and conforms to the required JSON contract.
derive_mf rows for the fixed set of malfunction fields (functional loss, over/under magnitude, early/late timing, unintended activation, stiction, direction error).nan vs non-nan consistency through slice outputs and subsequent Stage1R review.output/<RUN_ID>_stage1_derive_mf.json for downstream semantic review and later stages.Run the Stage 1 context preparation (only if needed) and then generate slice outputs per Function_ID to proceed into Stage1R semantic review before merging into the final Stage 1 derive file.
High-intent search queries and answers about installing and using this skill.
To derive function-level malfunction candidates, this Skill generates per-function rows for a fixed set of fields like functional loss, over/under magnitude, and timing errors from Stage 0 function mappings. It outputs a traceable JSON file ready for downstream semantic review.
The JSON schema contract enforces strict row counts, numbering rules, and mandatory nan versus non-nan consistency for malfunction fields. It ensures every output row is traceable to the original function detail_text and conforms to deterministic validation rules.
You generate Stage 1 malfunction entries by using Stage 0 function-to-ID mappings as input, producing slice outputs per Function_ID. These slices proceed into semantic review before merging into the final derive_mf JSON file.
No, this HARA Skill does not generate vehicle hazards, SEC ratings, scenarios, ASIL synchronization, or safety goals. It strictly isolates Stage 1 malfunction derivation, keeping risk and scenario knowledge out of this stage.
The fixed malfunction fields include functional loss, over magnitude, under magnitude, early timing, late timing, unintended activation, stiction, and direction error. Each function receives derived rows for all these applicable malfunction types.
Yes, you can automate functional safety malfunction derivation using pipeline orchestration. The Skill applies deterministic, contract-first workflows with controlled knowledge loading to produce consistent JSON outputs for pipeline integration.
WSJ40Extract HARA function mappings from 功能清单 tables into structured JSON.
WSJ40Convert Stage 1 safety faults into vehicle-level hazard JSON entries per Function_ID.
RoboSafe-LabAnalyze autonomous driving scenarios and define Operational Design Domains (ODD).
WSJ40Orchestrate Stage0–Stage4 HARA workflows with isolated sub-agents and validation gates.