Closure Retrospective
Use stable work evidence to decide whether durable guidance should be retained, removed, merged, clarified, scoped, relocated, or supplemented. Improve future decisions at justified context and maintenance cost. Generalize the decision method, not every resulting rule.
Boundaries
- Default to non-trivial task closure. An explicit retrospective or context review may examine a stable work unit even when other work remains open.
- Use existing evidence. Do not expand the task, change evidence access, or run new domain experiments to justify a lesson.
- Anchor the review in the current task. When necessary, consult only relevant, traceable prior evidence within existing access and task scope; distinguish its source and applicability. Do not default to broad history searches.
- Unresolved diagnosis limits only conclusions that depend on it. This skill does not resume or take over unfinished work.
- This skill governs durable guidance, not conversation compression or memory. Do not route output into memory workflows unless the user explicitly asks.
- Propose guidance edits before applying them. Explicit approval of the proposed scope authorizes those edits and appropriate verification; reuse that approval while it remains applicable. Invoking a retrospective alone does not authorize edits.
Workflow
1. Identify a context problem
Look for evidence that context influenced a decision that could recur. Common signals include missing, repeated, conflicting, stale, overly broad, or poorly timed guidance. Ask what helped, what obstructed, and what was missing.
For each candidate, connect the observation, the suspected or demonstrated contribution of context, and a specific future decision that a change would improve. A failed outcome, tool limitation, or execution mistake alone does not justify more instructions.
Completion criterion: every surviving candidate has evidence and a concrete decision improvement; otherwise use the no-change outcome in step 5.
2. Inspect existing coverage
Read the exact relevant guidance, its authority, and when it is available to the agent. Distinguish a content gap from a scope, discoverability, or execution problem before proposing a remedy.
Retain effective guidance. Existing coverage rules out redundant additions, but does not by itself rule out cleanup of duplication, conflict, or stale content. Identify the authoritative content to adjust; drop changes unlikely to improve the decision.
Completion criterion: every surviving candidate names an evidenced context problem and its existing coverage, including absence of a suitable artifact when verified.
3. Calibrate scope and choose the smallest change
Read references/decision-rubric.md for surviving candidates. Determine supported scope and loading time before choosing the artifact and edit. Deletion, consolidation, relocation, and additions are all valid outcomes.
Completion criterion: each proposal has bounded wording, an exact target and change, a justified benefit over its costs, and a reconsideration condition when it depends on changeable assumptions.
4. Resolve material uncertainty
After coverage and scope are clear, decide whether independent review could materially change an unresolved, consequential decision. Read references/independent-review.md only then; routine retention, relocation, or clarification normally needs no such review.
Dispatch subagents only when the user explicitly asks for subagent review or standing project instructions authorize it. Record review need, authorization, and availability separately. Evidence uncertainty limits conclusions; unavailable or unauthorized review is a process status, not evidence against an established fact.
Completion criterion: each proposal is ready, pending required review, or dropped. Narrowing may resolve uncertainty, but cannot waive an independently required review.
5. Present the decision
When nothing qualifies, finish normally; if the user explicitly requested the review, briefly explain why no change is warranted.
For surviving proposals or pending review, read references/output-template.md. Present enough detail to review the proposed change and determine the next authorized action. Keep proposed guidance, applied edits, and completed verification distinct.
Completion criterion: the user can assess the concrete change, its boundaries and costs, and any unresolved decision without reading the internal checklist.