Back to directory

Awesome Goal Prompts — Coding Agent Rescue Contracts

A searchable catalog of more than 300 runnable "/goal" contracts that give a coding agent a narrow, verifiable task with explicit constraints, a definition of done, and stop rules for uncertainty.

Awesome Goal Prompts — Coding Agent Rescue Contracts

The catalog holds runnable /goal contracts for engineering work: each one states a single goal, enough context to inspect, hard constraints on what not to touch, a definition of done, verification steps, and stop rules for uncertainty or risk. A developer copies the full contract block, including its GOAL, CONTEXT, CONSTRAINTS, DONE WHEN, VERIFY and STOP RULES sections, and pastes it to an agent such as Claude Code, Codex or Cursor.

The public catalog starts with 159 examples with public provenance drawn from official docs, GitHub threads, tutorials and tool READMEs, and keeps 200 additional reusable seed patterns in a separate section so they do not dilute the source-backed examples. A contribution guide and a companion skill support adding a new contract by editing a source file and regenerating the catalog.

More in Agent skills

See all 581
  • Provenance gate

    github.com

    A deterministic authorization gateway that sits between an AI agent and its tools, tracing where each parameter of a side-effect call came from, and blocking the call or asking a human when the source does not check out.

  • A markdown framework of system rules that restates Isaac Asimov's Four Laws of Robotics as a set of prioritized loss calculations for large-language-model agents to follow, published as a discovery file for agents and crawlers to read.

  • Vitruvius

    github.com

    An engineering research agent that runs a discover, read, synthesize, verify and review loop over technical questions, reading sources directly rather than inferring from titles, and publishing its own scored verifier benchmark, weaknesses included.