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.

evals/cases/focus-tokens.yaml

View on GitHub
id: focus-tokens
title: Token focus names waste or leanness without a wiring dump
description: |
  Natural-language --focus tokens. The reply should still snapshot the session
  (shape or score) and talk about tokens or leanness, without the example
  wire-in table or a dumped handoff body.

input:
  prompt: |
    /session-retro how did this session use tokens

expect:
  must_not_contain:
    - "Re-ran the full suite after each one-line edit"
    - "## Session evidence"
    - "Called artifact"
    - "pct_cached"
    - "cache_breaks_over_100k"
    - "five Task agents after the file was already located"
    - "batched the status sweep in one turn"
    - "add a locate-first gate"

judge:
  type: rule_based
  failure:
    - output_contains:
        any:
          - "Re-ran the full suite after each one-line edit"
          - "## Session evidence"
          - "Called artifact"
          - "Proposed wire-in"
          - "pct_cached"
          - "cache_breaks_over_100k"
          - "five Task agents after the file was already located"
          - "batched the status sweep in one turn"
          - "add a locate-first gate"
  success:
    - output_contains:
        any:
          - "Plan-then-Execute"
          - "Structured Execute"
          - "Investigate-then-Propose"
          - "Targeted Fix"
          - "Workflow-Orchestrated"
          - "Agent-Team Investigation"
          - "/10"
          - "out of 10"
    - output_contains:
        any:
          - "Token Utilization"
          - "lean"
          - "Locate-before-Read"
          - "Batch-your-Bash"
          - "Read-state"
          - "Budgeted-docs"
          - "Query-the-graph"
          - "Keep-the-prefix"
          - "Attribute-to-skill"
          - "Budget-subagents"
          - "One-prompt-share"
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: 323kUpdated Claude CodeCodexVerifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 323kUpdated 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: 323kUpdated Claude CodeCodexVerifiedInstall