agentleFS
KrishP147KrishP147/skills/find-me-a-skill/SKILL.md

find-me-a-skill

Skill37 starsChanged 47 days agoMITView on GitHub

Suggest which skill or workflow fits a task, with model and /effort per step and paste-ready commands; runs nothing. Use for "which skill should I use", "find me a skill", "how should I run this", "what model/effort for this".

The file

We split this skill into the rules it gives an agent

3 rules. About 0 are common, 2 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    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:

    Read-only: suggest, never invoke a skill, agent or tool that changes anything.

    line 8
  2. Distinctive:

    No repo or no board: ask what the task is, then stop.

    line 16
  3. Very specific:

    Agents: run inside a skill (pair, meta-orchestrator) or on request ("use the planner agent").

    line 28

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “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

Read-only: suggest, never invoke a skill, agent or tool that changes anything. The user starts the work by pasting your commands.

1. Task

First that applies:

  • Text after the command.
  • The goal of this conversation so far.
  • The top 1-3 board tasks, found as the next skill's BOARD.md says (~/.claude/skills/next/ or, as the plugin, kanban/next/); give one suggestion per task. No next skill installed: skip this line.
  • No repo or no board: ask what the task is, then stop.

2. Options (at most 3)

Candidates: every skill in your available-skills list, plus the entries and workflows in CATALOG.md.

  • Primary: the best fit among all installed skills, or a CATALOG.md workflow when the task spans its steps.
  • From your collection: label any option that comes from CATALOG.md, primary or not. A CATALOG.md entry or workflow that fits and isn't already the primary is added as another option.
  • Not installed: a CATALOG.md entry that is not in your available-skills list (a workflow: any of its skills) is labelled not installed only after a file check finds it missing: ~/.claude/skills/<name>/SKILL.md, then the krish plugin's install dir (its installPath in ~/.claude/plugins/installed_plugins.json, searched for <name>/SKILL.md; no such file or no krish entry → skip that check and say so). User-only skills never appear in that list even when installed. Give it the install command from the top README's Install section (./scripts/install-skills.sh, or /plugin install krish@krish). Its commands won't run until then.
  • Model and effort per step: CATALOG.md's default; for a large or risky task move effort up one level; move the model up one tier only when effort is already max. For a small task do the reverse: effort down one level, model down only when effort is already low. At opus with effort max (or haiku with low) there is nothing further; say so. Scale: haiku < sonnet < opus; low < medium < high < xhigh < max.
  • Not in CATALOG.md: name its origin (plugin <name> from its <name>: prefix, else built-in) and label model and effort estimate.
  • User-only skills (skill (user only) in CATALOG.md) are still suggested: the user runs them by pasting the command.
  • Agents run inside a skill (pair, meta-orchestrator) or on request ("use the planner agent"); hooks are installed, not run: say which.
  • No fit: say so, name the nearest option, and suggest doing the task directly at an estimate model and effort.

3. Reply

Task: <one line>

1. <name> (<skill|workflow>[, from your collection][, <origin>]): <why, one line>
   Steps: /a (sonnet · medium) -> you: <checkpoint> -> /b (opus · high)
   You decide: <each human-in-the-loop point>
2. ...

Run option 1:
/model sonnet
/effort medium
/a <args>

Steps and You decide come from CATALOG.md for a workflow; for a single skill, list its gates (approvals, yes/no points) from its description. Commands are exact and copyable; for a workflow give the first step's. Skills from the plugin (krish: entries in your available-skills list) are invoked with that prefix: write /krish:pair, not /pair.

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.