Stacks Writing For Agents

Writing for agents

Reference for every document an agent consumes in a Stacks project: a skill, the project AGENTS.md, a doc reached by a pointer. The packaging differs, the writing does not. The same levers make each one predictable, because the agent takes the same process every run rather than producing the same output.

Stacks ships 115 skills and expects projects to add their own under app/Skills/, so this is the skill that keeps that set from turning to sludge. The Stacks-specific mechanics (frontmatter, invocation, the override model, what buddy setup:ai does with the result) are in MECHANICS.md. Everything below is universal.

Credit: the model in this skill is adapted from Matt Pocock's writing-for-agents skill (MIT), https://github.com/mattpocock/skills.

Context pointers

A context pointer is a reference held in the agent's context that names some 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 a weakly worded pointer is a variance bug: sharpen the wording first, and 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 (a branch is a distinct case the document handles, so different runs take different paths through it). Every word of an always-loaded pointer costs on every turn, so it earns harder pruning than the body:

• Front-load the leading word. The pointer is where it does its triggering work.
• One trigger per branch. Synonyms that rename a single branch are one branch written twice. Collapse them, keep only genuinely distinct branches.
• Cut identity the body already carries. stacks-queue's description does not need to explain what a queue is.

The bundled skills follow one shape, and yours should too: `Use when

• , , . Covers and .The trailingCovers clause is doing pointer work, not decoration: it is how an agent holding a package name (@stacksjs/cache) or a path (config/queue.ts`) finds the skill that owns it.

The two loads

Every document and pointer you add spends one of two budgets:

• Context load is the cost of always-loaded material on the agent's window: an AGENTS.md line, a skill description, anything sitting in context every turn, spending tokens and attention 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, it is the price of human agency. Spend it where human judgement matters, remove it where it does not.

Material reached only through a pointer escapes context load at the price of the pointer's own line. Material with no pointer at all rides entirely on cognitive load.

Information hierarchy

A document is built from two content types: steps (the ordered actions the agent performs) and reference (definitions, rules, facts consulted on demand). The two mix freely: all steps (stacks-new-feature), all reference (stacks-orm), or both (stacks-investigate). The core decision is where each piece sits on the information hierarchy, a ladder ranked by how immediately the agent needs the material:

1. In-file step is the primary tier: what the agent does, in order.
2. In-file reference is consulted on demand. Often a legitimately flat peer-set (every rule of a review on one rung), which is a fine arrangement, not a smell.
3. Disclosed reference is pushed into a separate file, reached by a context pointer, loaded only when the pointer fires. Spans a sibling file in the same skill directory through fully external reference any document can point at.

Push too little down and the top bloats. Push too much and you hide material the agent actually needs. That tension is the whole decision.

Progressive disclosure is the move down the ladder so the top stays legible. Not primarily a token optimisation: it is how the hierarchy is protected. Branching is the cleanest disclosure test: inline what every branch needs, push behind a pointer what only some branches reach. stacks-technical-diagrams keeps 36 reference files beside one SKILL.md for exactly this reason.

Co-location is the within-file companion. Where the ladder decides how far down a piece sits, co-location decides what sits beside it once there. Keep a concept's definition, rules and caveats under one heading rather than scattered, so reading one part brings its neighbours with it. The test: the document should read like documentation written for the agent.

Sprawl is the failure mode: a document simply too long, even when every line is live and unique. Attention thins across the excess, and every extra line is one more to keep relevant. The cure is the ladder.

Steps and completion criteria

Every step ends on a completion criterion, the condition that tells the agent the work is done. Two properties make it a lever:

• Clarity. Can the agent tell done from not-done? A vague bound ("understanding reached") invites premature completion: ending the step before it is genuinely done, attention slipping to being done. The visible steps still ahead supply the pull, the criterion's clarity is the resistance. Defend in order: sharpen the bound first (local and cheap), and only if it is irreducibly fuzzy and you observe the rush, hide the later steps by splitting the sequence. Hiding only works across a real context boundary (a hand-off or a subagent dispatch). An inline call leaves the later steps in context and clears nothing.
• Demand. How much the criterion requires. "Every changed model accounted for" forces thorough work where "produce a change list" does not. Demand drives legwork, the digging the agent does within the work, latent in the wording rather than written as its own step. It is not step-bound: "every rule applied" binds a body of flat reference just as "every step done" binds a sequence, which is how an all-reference document still carries an exhaustiveness bar.

The strongest criteria are both checkable and exhaustive. stacks-investigate Phase 1 is the model to copy: one command, already run once, output shown.

When to split

Splitting one document into two spends one of the two loads, so split only when the cut earns it:

• By sequence. Split a run of steps where the post-completion steps tempt the agent to rush the one in front of it. Keeping them out of view drives more legwork on the current task. Beware the reverse: merging sequences exposes each step's later steps to what follows, inviting premature completion.
• By invocation. Skill-specific, see MECHANICS.md.

Leading words

A leading word is a compact concept already living in the model's pretraining that the agent thinks with while running the document (lesson, fog of war, tracer bullet, seam). Repeated as a token, never as a sentence, it accumulates a distributed definition and anchors a whole region of behaviour in the fewest tokens by recruiting priors the model already holds. Coining your own works if you define it clearly, but a made-up word recruits no priors: you pay in definition tokens what a pretrained word gives free. Reach for an existing word first.

It anchors twice. In the body, execution: the agent reaches for the same behaviour every time the word appears. In a pointer, invocation: when the same word lives in your prompts, your docs and your codebase, the agent links that shared language to the material and reaches it more reliably. This is why the Stacks skills insist on trait, driver, action, seam and tracer bullet rather than paraphrasing them.

Hunt for opportunities to refactor with leading words. A triad spelled out at three sites, a pointer spending a sentence to gesture at one idea. Each is a passage begging to collapse into a single token:

• "fast, deterministic, low-overhead" becomes tight (a tight loop).
• "a loop you believe in" becomes red, turning a fuzzy gate into a binary observable state (the loop goes red on the bug, or it does not).

Negation is the failure mode beside this lever. Steering by prohibition drags the forbidden behaviour into context and makes it more available, not less. Do not think of an elephant, and the elephant is all there is. Prompt the positive: state the target behaviour ("use signals and composables in stx templates") 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 positive target so attention lands on what to do.

Pruning

• Keep each meaning in a single source of truth: one authoritative place, so changing the behaviour is a one-place edit. Duplication costs maintenance and tokens, and inflates a meaning's prominence on the ladder past its real rank. It is the accidental inverse of a leading word, which repeats a token on purpose, never the meaning.
• The environment is a source of truth too, and in a Stacks project it is a rich one: buddy list, buddy <command> --help, config/*.ts, the storage/framework/*-auto-imports.json manifests, the generated storage/framework/types/*.d.ts. A document that restates it is a cache: a copy of a lookup, earning its load only when the lookup is expensive. Cache what the agent cannot find by looking (the unwritten convention, the reason behind a choice, the gotcha no config confesses) and leave the one-command lookups to the environment, where they cannot go stale. Every bundled skill's ## Gotchas section is this rule applied.
• Check every line for relevance: does it still bear on what the document does? A line loses relevance by never bearing on the task, or by going stale as the code it describes changes. Without a pruning discipline the default fate is sediment: stale layers that settle because adding feels safe and removing feels risky.
• Hunt no-ops sentence by sentence: an instruction the model already obeys by default pays load to say nothing. The test (does it change behaviour versus the default?) is model-relative, not reader-relative. Two people disagreeing about a no-op disagree about the default, and settle it by running the document, not by debate. When a sentence fails, delete the whole sentence rather than trim words from it. The test also grades leading words: a word too weak to beat the default (be thorough, when the agent is already thorough-ish) is a no-op, and the fix is a stronger word (relentless), not a different technique.

Before you finish

• Every line changes behaviour versus the default.
• Every meaning lives in exactly one place.
• The description names distinct branches, front-loads the trigger, and carries the Covers clause.
• Nothing in the body restates what buddy <command> --help or a config file already says.
• No em-dash anywhere in the file (the project-wide rule in AGENTS.md).
• bunx --bun pickier . is clean if you touched code alongside it.

Downstream

Reach for stacks-retro after a session to find which documents actually failed, and stacks-flow when the problem is that nobody remembers a skill exists.