Skip to content

renzrollon/shippable-skills / skills/productivity/session-retro

session-retro

Verified

Retro the current session. Score it against session shapes, flag token waste with next-turn fixes, and propose wire-ins into the skills this conversation actually called.

CHANGELOG.md

View on GitHub
# Changelog

## 0.2.0 (unreleased)

Handoff directory now follows the shared contract: a path you pass (`--handoff-
dir`) or name in the request, an instruction file override, then the first
existing of `.claude/handoff/`, `.agents/handoff/`, `docs/handoff/`, else
`.agents/handoff/`. The write location changes only in a repo whose sole handoff
directory is `docs/handoff/` (it now writes there instead of creating
`.agents/handoff/`), or when a path is passed or configured. Mentions of
`session-review` now read "a log-scoring tool such as agent-insights".

Token Utilization now scores the same dimensions Anthropic's `session-report`
looks for, without reading logs or inventing cache numbers.

- Findings first: 3-5 share-of-session lines (`most`, `~half`, `5×`, `lean`),
  then moments, then 1-4 recommendations that also feed Wiring.
- New tool-economy rules: Attribute-to-skill, Keep-the-prefix,
  Budget-subagents, One-prompt-share. Existing locate/batch/state rules stay.
- `--focus tokens` still forbids wiring tables and exact uncached / cache-read
  / output tables. Point at `session-report` for a 7-day numeric Claude Code
  report; this skill does not run it.

## 0.1.0 (2026-08-21)

First shipped release. Gate run the same day: `validate → evaluate → security → release`.

| Stage | Tool | Result |
| --- | --- | --- |
| 1 · validate | skill-lint 0.2.0 | passed, 0 findings (11 files) |
| 2 · evaluate | skill-up 0.7.0 | passed, 4/4 cases |
| 3 · security | SkillSpector 2.5.1 | passed, 0 actionable (1 reviewed suppress) |
| 4 · optimise | SkillHone | skipped (tool still being wired in SkillGantry) |
| 5 · release | skills 1.5.21 | passed, archive verified installable |

Evaluate cases, all PASS: `current-session-report`, `focus-tokens`, `focus-wiring`, `refuse-transcript-reread`.

The security suppress is SkillSpector rule P2 on an HTML comment in the wire-in template: visible author guidance, not a hidden agent directive. The baseline file stays local (gitignored).
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 CodeCodexUnverifiedInstall
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