skills/debrief/SKILL.md
Lightweight evidence-backed retro and catch-up reports for a current repo, branch, PR, backlog slice, or recent agent session. Use when the user asks for a debrief, catch me up, what changed, why it matters, product implications, end-user implications, developer experience implications, current app state, backlog state, workspace state, alternatives considered, or context rebuild after losing the thread. Trigger: /debrief.
npx skillsauth add phrazzld/agent-skills debriefInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
Produce a high-context explanation of what happened and why it matters. Default to read-only. This is a catch-up and sensemaking skill, not a delivery or codification skill.
Use /reflect instead when the main job is harness postmortem, operator
coaching, prompt debt, or writing follow-up codification. Use /demo when the
main job is a visual or runnable proof artifact. A debrief may point to either
as a next move.
Start from live evidence. Prefer:
git status --short --branch --untracked-files=allTreat memory and chat history as leads, not truth. If evidence is missing,
stale, or too expensive to refresh, say that in Sources & Gaps.
Do not imply validation without the exact command, path, route, artifact, or source that proves it. Separate confirmed facts, reasonable inference, and opinion.
Calibrate to the user's wording:
When using provider lanes, give them live artifacts only: diff, backlog, commit range, traces, or docs. Do not leak the lead's conclusions.
Use this shape by default. Compress sections when the change is small, but keep
Sources & Gaps, Current State, and Next Moves.
**Sources & Gaps**
- Confirmed sources:
- Missing or stale sources:
- Assumptions:
**Short Version**
One paragraph that catches up a busy operator.
**What Changed**
- User-visible or operator-visible behavior:
- Code/docs/tests/infrastructure:
- Validation evidence:
**Why It Matters**
- Product:
- End user:
- Developer experience:
- Codebase health:
**Decision Record**
- Decisions made:
- Alternatives considered or implied:
- Why those alternatives were not taken:
**Current State**
- Application/system:
- Backlog:
- Workspace/git:
**Risks & Unknowns**
- Confirmed risks:
- Unverified paths:
- Follow-up debt:
**Next Moves**
1. Immediate:
2. Next best:
3. Later:
Assume the reader has lost context. Define repo-local names the first time they matter. Do not dump raw diffs or command logs; synthesize them into an operator-useful story with citations to concrete files, commits, commands, or artifacts.
Make tradeoffs explicit. Good debriefs explain both "what we did" and "what we chose not to do." If alternatives are not visible in the evidence, say so instead of inventing a design debate.
Keep recommendations opinionated but bounded. The output should leave the user able to decide whether to ship, continue, pause, or reshape.
testing
Capture agent-session work records as local JSONL audit evidence. Links a backlog/spec, branch, commits, review verdicts, QA/demo evidence, transcript refs, and shipped ref without storing raw private transcripts. Use when: "trace this work", "write work record", "agent session trace", "journal this delivery", "link transcript evidence". Trigger: /trace, /journal.
data-ai
Turn proven agent-session patterns into first-party Harness Kit skills. Use when: "skillify this conversation", "make this into a skill", "generate a skill from current transcript", "extract reusable workflow". Trigger: /skillify.
testing
Run one targeted, read-only architecture or quality critique through a named lens from the shared rubric. Use when: "critique this module", "run an Ousterhout pass", "lens critique", "architecture critique". Trigger: /critique.
data-ai
Redact and package local agent-session excerpts for PRs, issues, receipts, or review evidence. Use when: "add agent transcript", "attach session proof", "show agent provenance", "redact transcript", "PR transcript". Trigger: /agent-transcript.