SECURITY REVIEW
Not yet assessed
Review the original instructions and requested permissions before installing.
No security review is available for this catalog entry yet.
Community / AGENT SKILL
causal-explanation
0
Prepare a substantive, source-backed causal account of an existing behavior, incident, or historical design decision when reconstructing and documenting the evidence is the selected deliverable. Do not trigger for routine rationale, explanation of your own recommendation, ordinary technical questions, current-task updates, or active failure diagnosis.
Understand the trade-offs.
SKILL QUALITY
Not yet assessed
How clearly the skill guides your agent, how complete its workflow is, and how you can check the outcome.
No quality assessment is available for this catalog entry yet.
The full skill.
Original instructions from the publisher’s SKILL.md
# Causal Explanation Answer the selected causal question with evidence calibrated to the claim. Keep observed facts, source-backed inference, competing explanations, and unknowns distinct. Do not turn an explanation request into implementation work. ## Route The Request - Bind the question to the exact behavior, decision, regression, threshold, or tradeoff. Infer the referent from current context only when the interpretation is safe; otherwise ask one targeted question. - Use `jcode` for a symbol lookup, file map, caller trace, or adjacent source context. Use the repo's QA intake or attribution lane for reproducing, diagnosing, or attributing an active failure, including a request to find an unknown root cause. An established cause alone does not select this skill; use it when reconstructing and documenting the causal account is the selected deliverable, or when the user explicitly invokes it. - Use the applicable architecture or performance skill when the requested outcome is a design review or benchmark judgment rather than an explanation. ## Build A Proportionate Evidence Record 1. Include firsthand user reports and relayed client updates as evidence; preserve their uncertainty and distinguish observations from causal claims. Do not demand independent corroboration merely because the user is the source. Identify the existing sources that can answer the question. Choose the cheapest authoritative evidence; pure historical recall uses memory first, while a current mechanism may be settled directly from source or an artifact. 2. Consult only sources likely to resolve the causal question. Use `jcode` for source structure and callers, `jdocs` for authored documentation or reference trees, direct QMD retrieval plus `get` or `multi_get` for exact prior wording or history, and `jdata` only when tabular evidence is material. Consult a relevant issue or review connector only when current evidence points there. Do not require an all-source sweep or enumerate and sweep connectors. 3. Record each source actually consulted. Mark relevant searches that returned nothing, unavailable sources that leave a material gap, and deliberately omitted categories whose evidence could not affect the answer. 4. For a regression, compare relevant successful and failed runs under like source/build, input, scenario, and environment conditions, and inspect the intervening changes. Completed qualification and later successes constrain the claim; timing is a hypothesis, not proof of cause. Code shape and runtime behavior can establish mechanism. They do not, by themselves, establish the motivation, rejected alternatives, or original intent. Attribute motivation only to an explicit source; otherwise label it as an inference or competing hypothesis. ## Calibrate The Explanation Separate the result into the smallest useful set of these categories: - **Direct observations:** current behavior or explicit source statements with precise citations. - **Causal inference:** the evidence chain connecting observations to the proposed explanation. State the confidence and why it is warranted. - **Alternatives:** explicitly rejected alternatives when the record says so; otherwise plausible competing hypotheses with evidence for and against. - **Gaps:** contradictions, empty searches, unavailable evidence, and questions the record cannot answer. Do not manufacture certainty. If the evidence proves mechanism but not motivation, say exactly that. Prefer a bounded `unproven` conclusion to a smooth story. ## Output Contract Lead with the supported answer or the fact that the cause is not proven. Then give the direct evidence, causal chain, material alternatives, gaps and confidence, and a compact sources-consulted record. Omit empty presentation sections when their absence cannot hide uncertainty. If this explanation belongs to an already selected implementation task, return the resulting preserve, change, avoid, and verification constraints to that same task. Do not create a ticket, next prompt, thread, worker plan, or successor task. Completion is terminal unless the user has already selected further work. For source and license details, read [references/provenance.md](references/provenance.md) only when auditing this skill's provenance.