Skip to content

TeamRetroHQ/teamretro-skills / skills/ai-session-retro

ai-session-retro

Verified

Use when a working session with an AI is wrapping up and should be reviewed and logged — "run a retro on this session", "AI session retro", "debrief this session", "log this session" — or when asked to retro a past session from a transcript. Offer it proactively at the natural end of a large or multi-step session. Do NOT use this to create or facilitate a retrospective in a retrospective product (TeamRetro or similar) — this skill reviews AI working sessions only.

evals/evals.json

View on GitHub
{
  "skill_name": "ai-session-retro",
  "evals": [
    {
      "id": 0,
      "eval_name": "fresh-capture-first-entry",
      "prompt": "That competitive analysis session is wrapped up (full transcript is in transcript-1.md). Before we finish — run a session retrospective on it and log it so the team can review AI-collaboration patterns later. There's no existing retro log yet.",
      "expected_output": "Creates docs/ai-retros/entries/<session-date>-<slug>.md containing one schema-compliant v1.2 entry: evidence-cited friction (audience/length not specified until turn 9, missing analytics access, stale Retrium pricing as agent-error), every friction item labeled from the fixed 10-label vocabulary AND ending with a '→ Fix:' line, a 'Do this first' section with exactly one fix, and no praise/filler. Entry is committed.",
      "files": ["fixtures/transcript-1.md"]
    },
    {
      "id": 1,
      "eval_name": "append-existing-entries-close-loop",
      "prompt": "Session's done — transcript-2.md has the full transcript of today's webinar email session. Run a retro and add it to our existing AI retro log (docs/ai-retros/entries/ in this repo). Please make sure past entries stay intact.",
      "expected_output": "Creates a new entry file without modifying the two existing v1.1-stamped entries; entry cites the merge-tag guess and missing Customer.io access with fixes; reads the most recent prior entry's 'One change' (legacy v1.1 section) and notes that confirming sequence structure upfront was adopted this session.",
      "files": ["fixtures/existing-entries/transcript-2.md", "fixtures/existing-entries/docs/ai-retros/entries/2026-07-06-support-ticket-theme-analysis.md", "fixtures/existing-entries/docs/ai-retros/entries/2026-07-09-onboarding-email-sequence.md"]
    },
    {
      "id": 2,
      "eval_name": "negative-trigger-product-retrospective",
      "prompt": "Can you create a retrospective for the design team for Friday? Start-stop-continue format, and add an icebreaker.",
      "expected_output": "Does NOT invoke ai-session-retro and does NOT write anything to docs/ai-retros/. Treats this as a product/facilitation request (e.g. TeamRetro retrospective creation or a clarifying question) — the skill's exclusion rule for retrospective-product work is respected.",
      "files": []
    },
    {
      "id": 3,
      "eval_name": "retroactive-backfill-transcript",
      "prompt": "Here's a transcript of an AI session from three weeks ago (transcript-1.md, session ran 2026-06-20). Backfill a retro entry for it into docs/ai-retros/entries/ — we're seeding the log with past sessions.",
      "expected_output": "Entry is dated 2026-06-20 (the session date, not today), filename starts 2026-06-20, header includes '**Backfilled:** yes', and the Step 4 close-the-loop check is skipped (no adoption claims about other entries).",
      "files": ["fixtures/transcript-1.md"]
    }
  ]
}
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