agentleFS
Sign inStart free
WellsonrWellsonr/skillset/skills/router/SKILL.md

router

Skill1 starsChanged 18 days agoMITView on GitHub

Shared routing procedure for all skillset router commands (make/development/plan/execute/review) — anti-stale pre-flight, task sizing, lean cross-plugin skill chains, per-phase model delegation, propose-then-confirm, rules application.

17 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use research and setup.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • NVIDIASkill

    research

    Before designing a feature, research how comparable libraries solve the problem, check best practices, known pitfalls, and upstream changes. Use when starting a new feature, investigating an unfamiliar area, or the user asks to research something.

    Also in 3 projects that use this

  • VercelSkill

    setup

    Set up Vercel CLI and project configuration. Use when the user says "set up Vercel", "configure Vercel", "link to Vercel", or "vercel init".

    Also in 3 projects that use this

The file

We split this skill into the rules it gives an agent

4 rules. 0 are common, 2 are distinctive, and 2 are very specific to this project.

  • 0

    Common

    Most of the 13 edited versions we read keep these.

  • 2

    Distinctive

    This author's own choices. Worth a look.

  • 2

    Very specific

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

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

    Anti-stale pre-flight: Run node "${CLAUDE_PLUGIN_ROOT}/scripts/scan.mjs" --check.

    line 10
  2. Distinctive:

    Detect trivial tasks: If the task is a small question or one-line request, say no chain is needed, answer directly, and stop.

    line 13
  3. Very specific:

    Selective load: Run node "${CLAUDE_PLUGIN_ROOT}/scripts/query.mjs" <task domain keywords> (top 5, short descriptions.

    line 14
  4. Distinctive:

    Never size below medium when the task touches auth, payments, DB migrations, or secrets.

    line 19

“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

Skillset Router Procedure

The calling command provides: an archetype (full starting chain) and a task. If the task contains --full, strip it and skip sizing (step 4). Follow these steps:

  1. Anti-stale pre-flight. Run node "${CLAUDE_PLUGIN_ROOT}/scripts/scan.mjs" --check.
    • Exit 1 with missing-catalog → offer /skillset:init first; do not continue until the user chooses.
    • Exit 1 with drift (added/version/removed) → offer /skillset:update; the user may decline and continue.
  2. Detect trivial tasks. If the task is a small question or one-line request, say no chain is needed, answer directly, and stop.
  3. Selective load. Run node "${CLAUDE_PLUGIN_ROOT}/scripts/query.mjs" <task domain keywords> (top 5, short descriptions; add --full only if nothing fits). NEVER read the full catalog.json/catalog.md. Then Read ~/.claude/skillset/data/rules.md; if missing → offer /skillset:init.
  4. Size the task (make/development only; plan/execute/review keep their archetype):
    • small (1–3 files, clear change): test-driven-development → verification-before-completion (+ commit-push-pr for development).
    • medium (multi-file feature): brainstorming → writing-plans → TDD → code-review → verification-before-completion (+ commit for development).
    • large (new project, cross-cutting): the full archetype chain.
    • Never size below medium when the task touches auth, payments, DB migrations, or secrets.
  5. Compose the chain from the sized chain, using only skills present in the catalog. Substitute both ways when unavailable: superpowers:test-driven-development ⇄ mattpocock-skills:tdd; pr-review-toolkit:code-reviewer ⇄ ecc:code-reviewer; mattpocock-skills:code-review ⇄ pr-review-toolkit:review-pr. No substitute → propose skipping with a note.
  6. Active rules. The always section is always active (selalu is a legacy alias); other sections (e.g. web-project) are active when their trigger matches the task.
  7. Assign a model per phase from the ## models section of rules.md (skill: haiku|sonnet|opus|main; main = stay in the session model). Missing section → haiku for verification, commit, finishing, doc sync; sonnet for implementation and reviews; main for everything else. Any skill that asks the user questions or needs approval ALWAYS stays main.
  8. PROPOSE → wait for explicit confirmation. Show the size and a compact table (skill | model | reason in ≤8 words), the active rules, and which archetype steps were trimmed. Execute nothing before the user approves (or edits, then approves).
  9. Execute. main phases: invoke via the Skill tool in order. Delegated phases: spawn an Agent with the chosen model and a self-contained prompt (goal, files, "invoke <skill> with the Skill tool, ask no questions, reply in ≤10 lines"). Delegate only mechanical, self-contained phases; otherwise run inline. Check relevant rules before each phase and respect each skill's internal gates.
  10. Final summary. Results per phase, rules applied, what was trimmed and why, follow-ups (/skillset:review, /skillset:update, --full to rerun with the whole chain).

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.

Check your setup, free