Skip to content
Agentmart
Submit a tool

retrospective

End-of-session learnings loop.

Skill by PlatformRelay

IndexedMIT★ 1
Write a review

Chalk Labs

Need help deploying retrospective?

Chalk Labs sets up agents like this inside real teams: connected to your tools, tested on your data, with a human in the loop where it matters.

Book a call

Install

Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.

npx agentmarket install skill/platformrelay-retrospective

Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.

Get the weekly diff

The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.

We use your email only for this newsletter. Privacy

About

End-of-session learnings loop. Extract what worked, what failed, and what was newly discovered, then write it back into agent-context and propose concrete skill/guideline refinements so the harness compounds over time. Use at the end of a session, after a lane lands, or when the operator says "retro / capture learnings / what did we learn".

Use cases: writing, agent-building

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.

Tools
No tools found
Secrets
None asked for
Install scripts
None found

Network hosts it talks to 1

  • codeload.github.com · found in install command

Files it touches

  • ~/.claude/skills/retrospective

What's inside

Skill name
retrospective
Description
End-of-session learnings loop. Extract what worked, what failed, and what was newly discovered, then write it back into agent-context and propose concrete skill/guideline refinements so the harness compounds over time. Use at the end of a session, after a lane lands, or when the operator says "retro / capture learnings / what did we learn".
License
MIT
Instructions
35 lines in SKILL.md · view on GitHub

Files (1)

  • SKILL.md

SKILL.md (first lines)

# retrospective — capture learnings, improve the harness

Turns a finished session into durable improvement. The harness should get better every time it runs;
this skill is how. Runs in the main session (it needs the session's own history) — not a subagent.

## Step 1 — reconstruct
Review what this session actually did: the lane(s), what landed, what got stuck, surprises, dead ends,
and any place a skill/guideline was wrong, missing, or slowed you down.

## Step 2 — extract (be specific, not generic)
- **Worked** — patterns/commands/approaches to keep. (Not "TDD is good" — "envtest `Eventually` with
  2s poll flaked; 5s fixed it".)
- **Failed / friction** — what wasted time; the root cause, not the symptom.
- **Newly discovered** — facts about the codebase/tooling not written down anywhere (a gotcha, a gate
  quirk, a hidden dependency).

## Step 3 — write it back
- Update `agent-context/coordination/SESSION-HANDOFF.md` (repo state, in-flight, next entry points).
- Durable project facts → the relevant repo's `agent-context/` (not here) or an ADR if it's a
  decision.
- Cross-session harness facts → propose a concrete edit to the specific skill/guideline that was
  wrong or missing, and (on confirmation) apply it. This is the self-improving loop — name the file
  and the change, don't just note "improve the skill".

## Step 4 — surface follow-ups & tidy the inbox
Anything discovered that deserves work → propose as backlog stories via `/write-story`. Then reconcile
the **INBOX**: tick or move to *Resolved* anything handled this session, and add any new open
question/report/PR the session produced, so the operator's review queue reflects reality.

## Do not
- Write vague, generic lessons — if it isn't specific enough to act on, cut it.
- Duplicate a fact the code or git history already records — capture what was non-obvious.
- Silently rewrite a skill — name the file and change, confirm, then apply.

From PlatformRelay/Kaddy, MIT.

More from PlatformRelay/Kaddy

Signals

Stars
1
Forks
0
Open issues
0
Last commit (this folder)
Jul 20, 2026 (3 months ago)
Last release
Jul 21, 2026 (3 months ago)

From GitHub (fetched Oct 24, 2026).

Details

License
MIT
First seen
Oct 24, 2026
Last checked
Oct 24, 2026

More from PlatformRelay

Done for you

Rather not set it up yourself?

Chalk Labs scopes, builds and runs agents for your team: the task audit, the eval harness, the integrations and the weekly ops report. Fixed scope, priced before we start.

See how we build AI agents →Read the case studies

Chalk Labs is a two-founder AI and growth studio. We build AI agents, CRMs and MVPs for Web3 and AI companies, and have shipped 50+ product and token launches.

launches shipped
50+
launches shipped
media placements
500+
media placements
founders on every account
2
founders on every account

Press we've placed: Forbes · CoinDesk · Decrypt · Cointelegraph · The Block

Rahil Jain, Co-founder, Chalk Labs. Ex-Goldman Sachs engineer. You talk to the person who builds it.

Get retrospective running in your business

Tell us what you want to automate. A founder replies within one working day with a plan, not a pitch deck.

Goes to the Chalk Labs founders, not the newsletter. Privacy

Need help deploying retrospective?

CallGet help
Cite this page

retrospective skill. Agent Mart. https://agentmart.dev/skills/platformrelay-retrospective (last updated 2026-10-24)