agentleFS
Infinite-Labs-AIInfinite-Labs-AI/know-your-unknowns-skills/know-your-unknowns/SKILL.md

know-your-unknowns

Skill2 starsChanged 4 months agoMITView on GitHub

Route and execute unknown-discovery workflows before, during, or after implementation. Use when a user asks for a blindspot pass, unknown unknowns, "what am I missing", domain vocabulary bootstrapping, design directions, mock-before-wire artifacts, codebase-grounded brainstorms, one-question-at-a-time interviews, reference semantics maps, tweakable plans, implementation notes, buy-in docs, merge quizzes, or interactive review artifacts/HTML artifacts that expose assumptions and capture human decisions.

3 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

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

  • 0

    Common

    Most of the 3 edited versions we read keep these.

  • 5

    Distinctive

    This author's own choices. Worth a look.

  • 1

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use the smallest workflow that resolves the unknown:

    line 16
  2. Distinctive:

    If more than one applies, sequence by phase: pre-implementation discovery, during-implementation logging, post-implementation …

    line 30
  3. Distinctive:

    Every unknowns artifact must include:

    line 34
  4. Distinctive:

    Do not create decorative artifacts.

    line 43
  5. Distinctive:

    When an interactive review artifact is useful, prefer a single self-contained HTML file with no build step.

    line 47
  6. Very specific:

    Read references/artifact-output-contracts.md for required sections by artifact type.

    line 49

“Common” is measured: kept by at least 6 in 10 of the 3 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

Know Your Unknowns

Purpose

Expose hidden assumptions before they become implementation mistakes. Treat artifacts as decision instruments: they should reveal uncertainty, let the human react, and turn that reaction into a better prompt, plan, sign-off, or implementation constraint.

This skill routes to the relevant leaf skill. If the user explicitly names one of the leaf skills, load that skill directly.

Decision Tree

Use the smallest workflow that resolves the unknown:

  • Unfamiliar code/module, risky change, "what am I missing": use unknowns-blindspot-pass.
  • Unfamiliar domain or weak vocabulary: use unknowns-domain-bootstrap.
  • User cannot specify taste or wants options to react to: use unknowns-design-directions.
  • UI/interaction is unclear and should be felt before wiring: use unknowns-mock-before-wire.
  • Problem is known but intervention space is not: use unknowns-intervention-brainstorm.
  • Requirements are ambiguous: use unknowns-spec-interview.
  • Existing implementation/reference must be ported or matched: use unknowns-reference-map.
  • User wants a plan but mostly needs to inspect judgment calls: use unknowns-tweakable-plan.
  • Implementation is underway and surprises/deviations must be preserved: use unknowns-implementation-notes.
  • Work is done but buy-in/sign-off is uncertain: use unknowns-buy-in-doc.
  • User wants to prove they understand a diff/change before merge: use unknowns-merge-quiz.

If more than one applies, sequence by phase: pre-implementation discovery, during-implementation logging, post-implementation buy-in/comprehension.

Universal Contract

Every unknowns artifact must include:

  • The user's original ask, reframed as the unknown being tested.
  • Evidence gathered, with clear separation between fact, inference, and assumption.
  • The hidden decisions or blindspots that matter.
  • A human action: choose, veto, answer, confirm, correct, sign off, or pass a quiz.
  • A copyable next prompt, decision summary, or implementation constraint.
  • A stop condition: what must be true before implementation/merge/sign-off proceeds.

Do not create decorative artifacts. If HTML is used, it must make review easier through layout, comparison, controls, state, or copy-back loops.

Artifact Mode

When an interactive review artifact is useful, prefer a single self-contained HTML file with no build step. Use it for side-by-side alternatives, selectable chips, toggles, quizzes, timelines, semantic maps, decision tables, or copyable follow-up prompts.

Read references/artifact-output-contracts.md for required sections by artifact type. Read references/html-artifact-guidelines.md before creating HTML.

Optional scaffold:

python3 ~/.codex/skills/know-your-unknowns/scripts/scaffold_unknowns_artifact.py --kind blindspot-pass --title "Auth SSO blindspot pass" --output /tmp/auth-blindspots.html

Source Map

This pack is modeled on Thariq Shihipar's "Know your unknowns" examples. For the page-by-page mapping, read references/corpus-map.md.

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.