agentleFS
victor-hac-workvictor-hac-work/ai-unknowns-resolver/skills/post/SKILL.md

post

Skill0 starsChanged 4 months agoMITView on GitHub

Post-implementation phase for a task run through unknowns:pre and unknowns:impl. Packages the work into a buy-in pitch doc, then generates a comprehension quiz the user must pass PERFECTLY before merge is recommended. Use after implementation is done, before opening/merging the PR, or when the user says "/unknowns:post", "pitch this", or "quiz me on the changes".

18 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

5 rules. 0 are common, 1 is distinctive, and 4 are very specific to this project.

  • 0

    Common

    Most of the 15 edited versions we read keep these.

  • 1

    Distinctive

    This author's own choices. Worth a look.

  • 4

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Write 09-pitch.md, then generate 09-pitch.html per ${CLAUDE_PLUGIN_ROOT}/skills/pre/references/design-system.md (§ 09-pitch.html).

    line 16
  2. Very specific:

    Write 5–8 questions that test understanding, not memory: "what happens if X fails?", "why option A over B?", "which existing code path …

    line 44
  3. Very specific:

    Write 10-quiz.md (report + questions + answer key at the bottom under ## Answer key) and generate 10-quiz.html per …

    line 45
  4. Very specific:

    Update ledger (Quiz score, Route status → complete if passed), regenerate index.html (quiz gauge green, MERGE-READY banner).

    line 51
  5. Distinctive:

    Not yet passed (user stops early):

    line 57

“Common” is measured: kept by at least 6 in 10 of the 15 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

unknowns:post — pitch, then prove you understand it

Two jobs: (1) get reviewers to buy in fast — they start with the same unknowns the author did; (2) make sure the ORCHESTRATOR actually understands what was built — a diff skim is not understanding, because behavior depends on existing code paths.

Step 0 — Locate state

Same resolution as unknowns:impl Step 1. Missing 08-notes.md is fine (zero deviations); missing directory entirely → offer to run in degraded mode on the current branch diff (create dir + minimal ledger first).

Step 1 — Pitch

Write 09-pitch.md, then generate 09-pitch.html per ${CLAUDE_PLUGIN_ROOT}/skills/pre/references/design-system.md (§ 09-pitch.html).

# Pitch: <task>

## Result
<demo first: what exists now, how to see it — screenshot/GIF pointer, URL, command>

## Problem
<2-3 sentences>

## Approach
<3-5 bullets, each tied to a ledger decision #>

## Unknowns we accounted for
| Reviewer would ask | We resolved it | Evidence |
<rows from resolved ledger entries a reviewer would have anticipated>

## Deviations from plan
<from 08-notes.md, honest, with why — or "none">

## Artifacts
ledger.md · 06-plan.md · 07-preview.html · 08-notes.md · index.html

Step 2 — Quiz (merge gate)

  1. Build the change report: context, intuition, what was done and why, touched code paths — reading the ACTUAL final code, not just the diff (behavior depends on existing paths the diff doesn't show).
  2. Write 5–8 questions that test understanding, not memory: "what happens if X fails?", "why option A over B?", "which existing code path handles Y now?". Mix multiple-choice and short-answer. Cover every deviation from 08-notes.md.
  3. Write 10-quiz.md (report + questions + answer key at the bottom under ## Answer key) and generate 10-quiz.html per ${CLAUDE_PLUGIN_ROOT}/skills/pre/references/design-system.md (§ 10-quiz.html — interactive, client-side scoring).
  4. Quiz the user IN CHAT, one question at a time (AskUserQuestion for multiple-choice). Grade against the key.
  5. Gate: perfect score required. Miss any → explain the missed area from the artifacts, then re-quiz ONLY the missed questions (rephrased). Repeat until perfect. Record final score + attempts in 10-quiz.md and the ledger.

Step 3 — Exit handoff (MANDATORY)

Update ledger (Quiz score, Route status → complete if passed), regenerate index.html (quiz gauge green, MERGE-READY banner).

  • Passed:

Quiz passed {score} (attempt {n}) — MERGE-READY. Pitch: 09-pitch.html — drop it in Slack / the PR description. Next: open the PR / merge. Optionally archive docs/unknowns-resolver/<slug>/ in the repo as decision history.

  • Not yet passed (user stops early):

Not merge-ready: {missed} area(s) unclear. Gaps: <list>. Re-run /unknowns:post to resume the quiz.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.