rca
RCA
Purpose
Build a branch-aware causal tree from apparent problem to defensible root causes. Classify each leaf by confidence. Produce concrete actions per cause.
Interaction Contract
- Ask one question at a time during interview phases.
- Keep the user informed when evidence changes branch direction.
- Confirm assumptions before promoting hypotheses to conclusions.
- If the user defers a question ("ask me later"), record the deferral, continue with the next question, and revisit before closing the interview phase.
- If the user contributes their own Q&A pairs, incorporate them into the tree exactly as you would an answer to your own question. Do not re-ask what the user has already answered.
Core Rules
- Interview first. No exceptions. Ask user questions to establish the baseline narrative before any codebase inspection. Do NOT open files, search code, or read logs until at least the five interview questions in Workflow §1 are answered.
- One question at a time. Do not comment on, interpret, or editorialize answers — record and ask the next question.
- Questions follow the incident narrative (what happened, why, what changed). Codebase observations can corroborate but must not hijack interview flow.
More from goblindegook/skills
grill-me
Use when user wants to stress-test an idea or plan, review a design, or says "grill me".
10tdd
Use when implementing behavior changes or bug fixes where automated tests can drive the implementation, or when asked to use TDD.
8roundtable
Review a project or feature from the point of view of a set of personas. Use this when the user asks for multi-perspective critique — "review from the POV of X, Y, Z", "what would a [role] think of this?", "roundtable review", "get different perspectives on this". Each persona reads the actual source code before commenting, then all personas discuss together to agree on a prioritised top 5 list. If no personas are provided, ask the user and offer suggestions.
4test-desiderata
Critiques and scores a project's test suite against Kent Beck's 12 test desiderata, assigning a 1–5 score to each property and surfacing the top 3 highest-impact improvements. Use this skill whenever the user wants to evaluate test quality, audit their test suite, understand what makes tests good, or get actionable feedback on how to improve their tests. Trigger on phrases like "review my tests", "how good are my tests", "test quality", "test audit", "improve my test suite", "evaluate my tests", "rate my tests", "Kent Beck", "test desiderata", or any request to assess or grade tests.
1