Documentation

Reference for writing any document an agent reads: a skill, an AGENTS.md, a docblock, a line in a README. The packaging differs; the levers do not.

Distilled from mattpocock/skills/writing-for-agents — see .agents/docs/DISTILLATION.md for what was kept, dropped, and added.

Context pointers

A context pointer is a reference held in the agent's context that names out-of-context material and encodes the condition for reaching it. A skill's description is one; a line in AGENTS.md naming a doc is the same object. The pointer's wording, not its target, decides when the agent reaches the material and how reliably. A must-have target behind weakly worded pointer is a variance bug: sharpen the wording first, inline the material only if sharpening fails.

A pointer does two jobs: state what the material is, and list the branches that should trigger reaching it. Every word of an always-loaded pointer costs on every turn:

• Front-load the leading word. The pointer is where it does its work.
• One trigger per branch. Synonyms renaming a single branch are one branch written twice; collapse them.
• Cut identity the body already carries.

The two loads

• Context load is always-loaded material on the agent's window: an AGENTS.md line, a skill description, anything spending tokens whether or not it fires.
• Cognitive load is the cost on the human: which documents exist and when to reach for each. The human is the index. Not a cost to minimise; spend it where human judgement matters.

Information hierarchy

Two content types: steps (ordered actions) and reference (facts consulted on demand). The core decision is where each piece sits:

1. In-file step — what the agent does, in order.
2. In-file reference — consulted on demand. Often a legitimately flat peer-set, which is a fine arrangement, not a smell.
3. Disclosed reference — pushed into a separate file behind a pointer.

Progressive disclosure is the move down the ladder. Branching is the cleanest test: inline what every branch needs, disclose what only some reach.

Co-location decides what sits beside what. Keep a concept's definition, rules and caveats under one heading. Grouped material reads as documentation; scattered material does not.

Sprawl is the failure mode: a document too long even when every line is live. The cure is the ladder.

Completion criteria

Every step ends on a completion criterion. Two properties make it a lever:

• Clarity — can the agent tell done from not-done? A vague bound invites premature completion. Sharpen the bound first, locally and cheaply.
• Demand — how much it requires. "Every modified model accounted for" forces thorough work where "produce a change list" does not.

The strongest criteria are both checkable and exhaustive.

Leading words

A leading word is a compact concept already in the model's pretraining that the agent thinks with: lesson, fog of war, tracer bullets. Repeated as a token, never as a sentence, it anchors a whole region of behaviour in the fewest tokens by recruiting priors the model already holds. A coined word recruits no priors: you pay in definition tokens what a pretrained word gives free.

Negation is the failure mode beside this lever. Steering by prohibition drags the forbidden behaviour into context and makes it more available. Don't think of an elephant. Prompt the positive: "write one-line comments" so the banned one is never spoken. A prohibition earns its place only as a hard guardrail you cannot phrase positively, and even then pair it with the target.

Pruning

• Single source of truth. Duplication costs maintenance and tokens, and inflates a meaning's prominence past its real rank. A leading word repeats a token on purpose; duplication repeats a meaning by accident.
• The environment is a source of truth too — package.json scripts, config, directory layout, --help. A document restating it is a cache, and earns its load only when the lookup is expensive. Cache the unwritten convention, the reason behind a choice, the gotcha no config confesses.
• Relevance, line by line. The default fate without this discipline is sediment: stale layers that settle because adding feels safe and removing feels risky.
• No-ops. An instruction the model already obeys by default pays load to say nothing. The test is model-relative, not reader-relative. When a sentence fails, delete the sentence rather than trim words from it.

In this fleet

• US English. User-facing prose belongs to the writing skill: it carries the WordPress Documentation Style Guide rules in references/ and hands the draft to humanizer or humaniseur-fr. This skill is for the documents an agent consumes.
• A docblock states what the code does, not what it should do. The fleet's most common documentation defect is a docblock that disagrees with the code — 'WordPress' where the function returns 'wordpress', a fixed array shape a failure path contradicts. WPCS checks that a docblock is present and shaped right; nothing checks that it is true. That is this skill's job.
• Do not restate the signature. @param int $id under function f( int $id ) is a no-op. Say what the value means, or say nothing.
• @return describes what comes back, including when it is nothing. "Empty string when the term has no archive" is worth a line; "the link" is not.