Skip to content
Skills Library
Personal productivity · Review & audit

aar

/aar

After Action Review of a chat that just ran. Answers the four AAR questions, keeps only concrete lessons, saves them to Claude memory and the vault, and files any skill or rule change as a proposed edit in the Switchboard to-do. Use when the user says "aar", "after action review", "review this session", "what did we learn" or "run an aar", when a handoff note says to run it on the old chat, and before the finishing standup.

Repository
switchboard ↗
Licence
Apache-2.0
Last repository change
10.11.2026

Assessment and security

90/ 100Class: good
Format25 / 25
Frontmatter with name and description, valid name, description up to 1,024 characters, name matching the folder.
Discoverability25 / 25
Description explains when to use the skill, has a reasonable length and uses third person.
Content25 / 35
At least 150 words and at most 500 lines, headings, step lists, examples or code.
Hygiene15 / 15
No placeholders such as TODO, no Windows paths, with a licence.
Riskno signals detected in the scan
Files in this package
2
Scripts
0
HTML comments
1
URLs in SKILL.md
0
Deprecation

No signals detected. Repository changed on 10.11.2026, without references to retired models or APIs.

Licence

Apache-2.0, otwarta: można używać i zmieniać, z zachowaniem licencji i informacji o zmianach.

What to check before installing
  • SKILL.md zawiera komentarze HTML (1). Nie widać ich w podglądzie Markdown, a model je czyta.

What it does and when to use it

After Action Review of a chat that just ran. Answers the four AAR questions, keeps only concrete lessons, saves them to Claude memory and the vault, and files any skill or rule change as a proposed edit in the Switchboard to-do. Use when the user says "aar", "after action review", "review this session", "what did we learn" or "run an aar", when a handoff note says to run it on the old chat, and before the finishing standup.

Triggers from the description

aarafter action reviewreview this sessionwhat did we learnrun an aar
Original author description

After Action Review of a chat that just ran. Answers the four AAR questions, keeps only concrete lessons, saves them to Claude memory and the vault, and files any skill or rule change as a proposed edit in the Switchboard to-do. Use when the user says "aar", "after action review", "review this session", "what did we learn" or "run an aar", when a handoff note says to run it on the old chat, and before the finishing standup.

How it works

  1. Which chat
  2. The four questions
  3. The lesson bar
  4. Where it goes
  5. Run it honestly
  6. Report
Original SKILL.md headings · 55 lines, 558 words
SKILL.md preview
---
name: aar
description: After Action Review of a chat that just ran. Answers the four AAR questions, keeps only concrete lessons, saves them to Claude memory and the vault, and files any skill or rule change as a proposed edit in the Switchboard to-do. Use when the user says "aar", "after action review", "review this session", "what did we learn" or "run an aar", when a handoff note says to run it on the old chat, and before the finishing standup.
---

# AAR

Debrief the chat that just ran, so the next one does not repeat its mistakes or forget what worked. Based on the Army's After Action Review; adapted from aar-loop by Cooper Simson (MIT, see LICENSE).

## Which chat

- Finishing: this chat.
- From a handoff note: the old chat. Read its transcript (the Old transcript path in the note) with grep and targeted reads, not whole. You are usually a subagent here; keep the report short.

## The four questions

Answer each in a line or two, plainly, about what the agent did. Not what the user probably thinks happened.

1. What was supposed to happen?
2. What actually happened?
3. Why was there a difference?
4. What do we do the same or differently next time?

Question 3 wants a mechanism. "The hub returns 304 with an empty body, so the parse threw" is a why. "Should have been more careful" is not.

No difference is a valid answer. Then question 4 locks in what worked: a pattern that held up is a lesson too.

## The lesson bar

A lesson ships only if another agent, months later, could act on it with no judgment call. It names a tool, file, command, number or condition.

Ship:
- "claude-mods has no package.json; run a mod's tests with `claude plugin test ./usage-meter`."
- "Editing extension/kit in switchboard is lost on the next build; edit claude-mods/kit instead."

Drop:
- "Be more careful with builds."
- "Check things before finishing."

Push a vague lesson until it is concrete, or drop it. Zero lessons is fine; do not invent one.

## Where it goes

1. **Memory**, one file per lesson, in this project's memory folder (`~/.claude/projects/<project slug>/memory/`). Read `MEMORY.md` there first; if a memory already covers it, update that file instead of adding one. Run `hub_recall` for the topic first. If the lesson replaces an older one, write the new note with `supersedes: <old-slug>` rather than duplicating or rewriting it; use `contradicts:` if unresolved, and add `related:` or `depends-on:` where they fit. New file: frontmatter with `name`, `description`, `metadata.type: feedback`; body is the rule, then `**Why:**` (what happened in this chat) and `**How to apply:**`. Add a one-line entry to `MEMORY.md`.
2. **Vault**: append the whole AAR (date, chat, four answers, lessons) to `~/Documents/vault/wiki/aar-<project>.md`. Follow `~/Documents/vault/README.md`: if the page is new, add `- [[aar-<project>]] ...` to `wiki/index.md`; append a dated line to `wiki/log.md` either way. Link memory files by `[[name]]`.
3. **Skill or rule changes are never applied.** If the next chat will repeat the mistake unless something it loads changes, write a fix plan: exact target file and exact text to add or replace. File it with `mcp__switchboard__todo_add` for the user to approve. Kit rules, skills, agents and mods live in `claude-mods` (`kit/rules.md`, `kit/skills/<name>/SKILL.md`, `kit/agents/<role>.md`), never the copies in `~/.claude` or `switchboard/extension/kit`.

Not every lesson needs a fix plan. Context only goes to memory and the vault.

## Run it honestly

An AAR that always says "went great" is useless. Find the real friction point, even on a run that mostly worked: the retry, the wrong guess, the test that was skipped. Do not dress up context as a fix plan.

## Report

The four answers, the lessons with their memory file paths, the vault page, and the to-do items filed (or "none").

<!-- switchboard generic copy -->
Files in this package (2)
  • LICENSE 1 KB
  • SKILL.md 4 KB

Versions

The database has no other skills named aar.

Database snapshot: 13.11.2026 · package version a4d5f2094711 (SHA-256 digest of the entire package)

Downloads in the last 30 days: 0 · ranking

Report a problem with this skill: harmful content, incorrect descriptions or an author’s removal request.

Similar skills

  • 100self-improve
    Use when the user says "self-improve", "/self-improve", "improve this skill", "tune the ops agent", Switchboard's Learn automatically run, or an aar or improve-system report has filed a proposed edit to a kit skill, role agent or rules.md section and it needs testing before the…
  • 100after-action-report
    Run a structured after-action review (postmortem, retrospective) on a launch, incident, or completed project to capture timeline, root cause analysis, contributing factors, and actionable lessons. Use this skill whenever the user wants to run a postmortem, retrospective, AAR, or…

Feedback and reports

Your message goes to Bartosz Sokoliński. An email address is needed only for a reply. See the privacy policy.