agentleFS
atpaawejatpaawej/opencode-skillpack/skills/clarity/SKILL.md

clarity

Skill0 starsChanged todayMITView on GitHub

Distill a tangled thought, stuck decision, or messy artifact into what is actually happening, the crux, hidden assumptions, and the next concrete move.

13 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

8 rules. 0 are common, 7 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 13 edited versions we read keep these.

  • 7

    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. Very specific:

    If the working directory has a GLOSSARY.md (or GLOSSARY-MAP.md), read it silently first and hold to its terms.

    line 11
  2. Distinctive:

    Plain, calm, direct prose: Every sentence earns its place.

    line 15
  3. Distinctive:

    Split conflated terms: When one vague word (*"broken"*, *"scale"*, *"risky"*, *"stuck"*) bundles two or three distinct ideas, name each …

    line 17
  4. Distinctive:

    When invoked with no context or only a bare signal (*"I'm stuck"*), skip explaining this framework and ask two short prompts:

    line 24
  5. Distinctive:

    Wait for the dump before moving to step 2.

    line 28
  6. Distinctive:

    Read the single reference matching the dominant shape of the input.

    line 32
  7. Distinctive:

    When the user points to a file or thread, read only what they pointed to.

    line 40
  8. Distinctive:

    Check before finishing: every conflated term split, every unverified constraint marked as an assumption, and one concrete move named.

    line 52

“Common” is measured: kept by at least 6 in 10 of the 13 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

Clarity

Cut through fog. Take whatever messy thought, stuck choice, or tangled artifact the user brings and surface what is real, what matters, and what moves next.

If the working directory has a GLOSSARY.md (or GLOSSARY-MAP.md), read it silently first and hold to its terms.

Voice and discipline

  • Plain, calm, direct prose. Every sentence earns its place.
  • Facts apart from assumptions. Label unverified claims, fears, and inferences as assumptions rather than treating them as fixed reality.
  • Split conflated terms. When one vague word ("broken", "scale", "risky", "stuck") bundles two or three distinct ideas, name each meaning separately.
  • Stay self-contained. End with a concrete move the user can take immediately; never assume external tools or other skills.

Working loop

1. Draw out the dump if empty

When invoked with no context or only a bare signal ("I'm stuck"), skip explaining this framework and ask two short prompts:

  1. What is happening right now? Dump it raw — half-baked or contradictory is fine.
  2. What decision, friction, or outcome is forcing the issue today?

Wait for the dump before moving to step 2.

2. Route to the fog's shape

Read the single reference matching the dominant shape of the input. Load a second only when the dump genuinely spans two shapes.

The dump looks likeRead
Overload, competing threads, mixed facts and worry, unclear bottlenecktangles
Circling options, false binaries, competing trade-offs, analysis paralysisdecisions
A pasted or referenced note, doc, thread, or half-formed concept with blurred pointsartifacts

When the user points to a file or thread, read only what they pointed to.

3. Distill through five lenses

Produce a crisp breakdown using the lenses that actually apply (skip any lens that would only force filler):

  1. What's actually happening — observable facts stripped of story, worry, or padding.
  2. The crux — the single question, tension, or bottleneck everything else hangs on.
  3. Hidden assumptions & false constraints — what is treated as fixed, binary, or urgent without evidence, plus any conflated terms split apart.
  4. Signal vs. noise — what moves the needle now vs. secondary threads to park.
  5. Next concrete move — the smallest reversible step or choice that resolves or tests the crux.

Check before finishing: every conflated term split, every unverified constraint marked as an assumption, and one concrete move named.

4. Pinpoint remaining unknowns

End with 1–3 sharp questions only when a critical unknown blocks naming the crux or next move. When the fog is already clear enough to act, stop after Next concrete move.

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.