Skip to content

borghei/Claude-Skills / workflow/session-retro

session-retro

Verified

Review a coding-agent session and turn its failures and waste into ranked fixes to the agent's environment. Use when a session went slow or wrong, after a messy agent run, or to tune CLAUDE.md or AGENTS.md.

assets/retro_template.md

View on GitHub
# Session Retro — <session or task name>

**Session:** <transcript path or session id>  ·  **Date:** <YYYY-MM-DD>  ·  **Reviewer:** <name>
**Task the session was given:** <one sentence>
**Outcome:** <shipped / partly shipped / abandoned> in <N> tool calls, <M> failed

## Analyzer summary

<Paste the text output of session_log_analyzer.py here. Delete detectors that fired on
legitimate work and say why in one line.>

## What cost the most

List the three moments that burned the most calls or time, with call numbers. One line each:
what the agent was trying to do, what it did instead, and what it lacked.

1. <calls #a-#b: ...>
2. <calls #c-#d: ...>
3. <calls #e-#f: ...>

## Action items (ranked, high first)

| rank | category | change | file | evidence | severity |
|------|----------|--------|------|----------|----------|
| 1 | missing-check | <imperative change> | <path/to/file> | <calls #n-#m: what happened> | high |
| 2 | navigation | <imperative change> | <path/to/file> | <calls #n: what happened> | medium |

Categories: navigation, missing-check, standards, steering-file, tool-economy,
permission-friction, skill-miss, information-access.

## Parked

Items that are real but not worth a change yet, each with the trigger that would promote it
(for example "promote if it recurs in a second session").

- <item> — promote when <trigger>

## Verification

How the next session proves each shipped item worked (re-run the same task, or watch for the
same call pattern in the next transcript).

- Item 1: <check>
reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness.Stars: 322kUpdated Claude CodeCodexVerifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 322kUpdated Claude CodeCodexVerifiedInstall
executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is availableStars: 322kUpdated Claude CodeCodexVerifiedInstall