Use this skill when closing a finished task — writing the SUMMARY.md final report, the phase-09 retrospective, and task documentation in the workspace task dir. Don't use it for mid-task status updates or commit messages.
Scanned 9/5/2026
Install to Claude Code
npx -y skills add atretyak1985/swarmery --skill session-closeout --agent claude-codeInstalls into .claude/skills of the current project.
Are you the author of Session Closeout?
Add the live security badge to your README — it updates automatically with every re-scan.
[](https://www.skillsdirectory.com/skills/atretyak1985-session-closeout)More formats (shields.io, HTML) on the badges page.
---
name: session-closeout
version: "1.0.0"
owner: "swarmery-core"
description: "Use this skill when closing a finished task — writing the SUMMARY.md final report, the phase-09 retrospective, and task documentation in the workspace task dir. Don't use it for mid-task status updates or commit messages."
color: yellow
docs:
status: draft
updated: 2026-09-01
---
# Purpose
Closing artifacts are how finished work becomes visible: the dashboard reads
`SUMMARY.md`, and the retro/advisor loop ingests `phases/09-retrospective.md`
and the delegation ledger. A task closed without them is invisible work.
# What to write
1. **`{task-dir}/SUMMARY.md`** (canonical final report, ≤300 lines) — exactly
these seven sections: `## Результат`, `## Змінені файли`, `## Агенти`,
`## Сесії`, `## Відхилення від плану`, `## Скріншоти`, `## Follow-ups`.
Deviations compare actual agents/loops against ORCHESTRATION.md, not only
the plan; the screenshots section lists `{task-dir}/screenshots/` files
with captions or states "None captured". Format details:
`resources/summary-format.md`.
2. **`{task-dir}/phases/09-retrospective.md`** (≤150 lines) — for non-trivial
tasks. MUST use the machine-parsed headings: `## Lessons Learned` with
`### Lesson N: <title>` subsections each carrying a `**Action**: …` line,
and `## Process Improvements` (that exact heading — the ingester regexes
match it, not synonyms). Read `{task-dir}/logs/agents.md` (7-cell ledger)
first — its quality/mistakes cells are the evidence base. Format:
`resources/retrospective-format.md`.
3. **Archive** when every criterion is met:
`bash "${CLAUDE_PLUGIN_ROOT}/bin/agent-work.sh" complete {task-id}`. Never
archive with unmet criteria or a missing SUMMARY.md — escalate instead.
Honesty rule: outcome is `Success | Partial | Failed` as it actually was;
lessons cite specific phases/files; a flattering retro poisons the
self-improvement loop.
# How to use
## What it does
Carries the closing-artifact contract for a finished task: the seven-section `SUMMARY.md`, the machine-parseable retrospective, and the archive step — the formats the dashboard and the retro/advisor loop actually ingest.
## When to use it
At the end of any orchestrated or plan-driven task: the last checkbox is ticked, the work is verified, and the task dir needs its final report and retrospective before archiving.
## How to invoke
`@core:tech-lead` and `@core:implementation-agent` load it via `skills:` at close-out; load it manually when finishing a task by hand and follow `resources/summary-format.md` + `resources/retrospective-format.md`.
## Worked example
After a 3-phase plan finishes, the orchestrator writes `SUMMARY.md` (result, changed files, agents used with loop counts, deviations: "phase 2 needed one correction loop — planned 0"), then `phases/09-retrospective.md` with `### Lesson 1: verify fixture DBs after migrations` and its `**Action**:` line under `## Lessons Learned`, one entry under `## Process Improvements`, and runs `agent-work.sh complete 2026-09-01-rate-limiting`.
Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.
No comments yet. Be the first to comment!