agentleFS
IgorGanapolskyIgorGanapolsky/mac-yolo-safeguards/.agents/skills/agentic-pattern-selector/SKILL.md

agentic-pattern-selector

Skill3 starsChanged 3 days agoMITView on GitHub

Select the minimum justified agentic patterns for a typed task and emit a deterministic evidence receipt.

The file

We split this skill into the rules it gives an agent

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

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 3

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use this before designing or expanding a non-trivial agent workflow.

    line 8
  2. Very specific:

    Start from assets/read-research.example.json or assets/external-write.example.json.

    line 12
  3. Very specific:

    Treat status=block as a hard stop.

    line 20
  4. Very specific:

    Use selected as the smallest justified architecture and gates as required evidence.

    line 21

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 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

Agentic Pattern Selector

Use this before designing or expanding a non-trivial agent workflow. It prevents pattern cargo-culting: every one of the 21 source patterns is either selected with a reason or rejected with a reason.

Workflow

  1. Start from assets/read-research.example.json or assets/external-write.example.json.
  2. Describe effects and risk explicitly. Never infer an external write from prose alone.
  3. Run the selector:
node tools/agentic-pattern-selector.js --manifest path/to/task.json
  1. Treat status=block as a hard stop. Fix the manifest; do not bypass it.
  2. Use selected as the smallest justified architecture and gates as required evidence.
  3. For a full system decision brief, wire the same receipt into the existing decision stack:
node tools/agent-decision-stack.js \
  --task "<measurable decision>" \
  --governance infra \
  --pattern-manifest path/to/task.json \
  --json

Hard rules

  • external_write, and internal_write with risk=high, require humanConfirmation=required.
  • Parallelization requires at least two independent workstreams and disjoint resources.
  • Multi-agent requires parallel eligibility plus at least two explicit specialist roles.
  • Retrieval work requires source provenance and unsupported-claim checks.
  • Simple deterministic work should normally select only evaluation/monitoring.
  • Credential-shaped field names or string values are rejected without echoing attacker-controlled content.
  • Input must be valid UTF-8; decoding is fatal rather than replacement-based.
  • Valid JSON hashes its canonical parsed value, so whitespace and key order do not change the receipt. Malformed encoding/JSON hashes the original bytes.
  • A receipt explains architecture; it does not prove execution. Execute every returned gate and retain its evidence separately.

Verification

node --test tests/test-agentic-pattern-selector.js
node tests/test-agent-decision-stack.js
bash .agents/skills/agentic-pattern-selector/test.sh

Sources

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.