AGENT SKILL

root-cause-debugging

Infer and test the smallest causal explanation instead of stacking speculative fixes.

License 许可
MIT
所属仓库
syr-cn/PluginRSI
最近更新
2026-10-18
Stars
★ 61
Forks
1

01 前言 Front Matter

name
root-cause-debugging
description
Infer and test the smallest causal explanation instead of stacking speculative fixes.

02 资源文件 4

03 SKILL.md 正文

Root-cause Debugging

Internal ID: S13
Category: diagnosis
Tier: core
Output: RootCauseDiagnosis

Activate when

  • The workflow reaches the diagnosis stage and the required inputs are present.

Applicable roles

  • Critic
  • Solver

Required inputs

  • issue_profile
  • reproduction_result
  • localized_code

Optional inputs

  • trajectory
  • memory_hits

Procedure

  1. Separate symptom, trigger, faulty state or assumption, and violated contract.
  2. Form two or more competing hypotheses when uncertainty is material.
  3. Choose the cheapest discriminating check.
  4. Update or reject hypotheses from fresh evidence.
  5. Stop when one mechanism explains the failure and predicts the fix.

Verification

  • The chosen cause explains both observed and expected behavior.
  • Rejected hypotheses retain their evidence.

Stop conditions

  • One falsifiable root cause is supported or diagnosis is explicitly inconclusive.

Constraints

  • Do not apply multiple speculative fixes at once.
  • Do not label the first suspicious line as root cause.

Allowed tools

  • read_range
  • shell_exec
  • inspect_failure
  • text_search

Provenance

Synthesized from design patterns in superpowers, addy-agent-skills, swe-agent. This file is a normalized reusable abstraction, not a verbatim copy of an upstream skill.

04 同仓库更多 Skills