Agent-Skills.md

Agent Skills: Determinism

Use when verifying outcomes with code instead of LLM judgment, versioning prompts with hashes, or ensuring reproducible agent behavior. Load for any critical verification. Scripts return boolean exit codes, not subjective assessments. Prompts use semantic versioning with SHA256 validation.

UncategorizedID: ingpoc/SKILLS/determinism

Repository

ingpoc/SKILLS
ingpoc
8

Install this agent skill to your local

pnpm dlx add-skill https://github.com/ingpoc/SKILLS/tree/HEAD/determinism

Skill Files

Browse the full folder contents for determinism.

Download Skill

Loading file tree…

determinism/SKILL.md

Skill Metadata

Name
determinism
Description
"Design machine-checkable assertions, stable hashes, fixtures, and exit-code contracts for behavior that must be reproducible. Use when an outcome currently depends on subjective agent judgment or when prompts and generated artifacts need versioned integrity checks. Do not use merely to run an existing repository verification gate; that is verify's job."

Determinism

Turn a repeated judgment into an executable contract.

Procedure

  1. Name the exact input, expected invariant, and authoritative readback.
  2. Use the simplest deterministic mechanism: parser, schema, fixture, checksum, query, or narrow test.
  3. Return a meaningful exit status: 0 for satisfied, nonzero for a specific failure class.
  4. Keep diagnostics compact and stable enough for another script or agent to consume.
  5. Version prompts or generated contracts when their semantics affect the result; record a hash when drift must be detected.
  6. Prove both a passing case and a deliberately failing case.

Boundary

  • determinism designs the assertion or integrity mechanism.
  • verify selects and runs the repository's declared gates.
  • testing skills construct broader behavioral test suites.

References