What problem does it solve? Documents written for AI agents often fail unpredictably: agents miss referenced material, rush steps, or ignore instructions. This Skill provides a systematic reference for writing agent-facing documents so the agent follows the same process every run, covering skills, AGENTS.md, CLAUDE.md, and any doc reached by a pointer. ## Core Features & Use Cases - Context pointer design: Rules for writing descriptions and pointer lines that reliably trigger the agent to reach out-of-context material, with branch-based trigger wording. - Information hierarchy and progressive disclosure: A ladder model (in-file steps, in-file reference, disclosed reference) for deciding what stays inline and what moves behind a pointer. - Completion criteria and splitting: Guidance on writing checkable, exhaustive step completion criteria and deciding when to split documents by sequence or invocation. - Use Case: When authoring a new skill, use this reference to write a frontmatter description with distinct trigger branches, prune no-op sentences, and decide between model-invoked and user-invoked invocation via the companion SKILL-MECHANICS.md. ## Quick Start Ask the agent to review your draft skill or AGENTS.md file using the writing-for-agents principles and suggest pruning, pointer, and structure improvements.
