write-docs
Skill42 starsChanged 12 days agoMITView on GitHub
Write, reorganize, or review end-user documentation, navigation, examples, and screenshots, including moving, renaming, merging or deleting a page (which needs a redirect). Ground instructions in the running product, choose structure for the reader's task, and verify the rendered result. Use for docs content and its structural checks; use write-translations as well for localized content.
43 other projects use a version of this.
Adopted by
Most recently taken up by
- alchemy-run/alchemy-effecttheir own version, first seen Nov 9, 2026
- alchemy-run/alchemytheir own version, first seen Nov 8, 2026
- adhamhaithameid/write-skillstheir own version, first seen Nov 8, 2026
- dzhng/jevgreptheir own version, first seen Nov 7, 2026
- oberon-systems/enggraphtheir own version, first seen Nov 6, 2026
- dzhng/skillstheir own version, first seen Nov 4, 2026
Used with
Projects with it also use close-spec, design-with-images and 10 more.
Other files people use alongside this one.
Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
dzhngSkillclose-spec
When a spec under specs/<feature>/ is done shipping (a write-spec build, or any planned task that produced a spec), archive it to specs/done/ and rewrite it from a build-plan into a durable rationale — the why, the principles, the invariants — pointing back to the real code for the how. Use when implementation has landed and the plan no longer matches what shipped, or the user says a feature/spec is finished. Pairs with [write-spec](../write-spec/SKILL.md) (the plan this closes) and [review](../review/SKILL.md) (run before closing).
Also in 5 projects that use this
dzhngSkilldesign-with-images
Explore UI and visual design with image generation, then iterate the real implementation against the user-selected concept. Use when brainstorming visual directions, offering mockup variants, or implementing an approved generated design; do not stop at the mockups or the first code approximation.
Also in 5 projects that use this
transitive-bullshitSkillcompare-screenshots
Compare screenshots against the intended design, distinguishing approved references from historical baselines. Use for iterative implementation matching, before/after visual review, objective image telemetry, or checking a lone capture for flat, empty, or badly framed content.
Also in 5 projects that use this
LucasMatuszewskiSkillaudit-agents
Audit and maintain agent configuration files (.claude/agents, AGENTS.md, CLAUDE.md, agent-memory). Checks for duplicates, outdated info, naming conflicts, wrong technical facts, and files over 200 lines. Fixes all issues and commits.
Also in 5 projects that use this
wolverin0Skillaudit-decisions
Standardize remediation decisions, baseline-policy entries, and legal-review banners for accepted, deferred, or policy-blocked audit findings.
Also in 4 projects that use this
rhnfzlSkillexplore-unknowns
Walk the user through mapping the unknowns of a task before, during, and after they build it, and hand them a four-quadrant map (settled ground, open decisions, taste, and landmines) they keep. Reach for this eagerly whenever a request is ambiguous or underspecified, the codebase or domain is unfamiliar, the user will "know it when they see it", a reference implementation must be understood before porting, a build-queue or architecture-changing task is about to start, a mid-build deviation needs capturing, or a finished change needs buy-in or verified understanding before merge. Trigger even when the skill is not named: "do a blindspot pass", "find my unknown unknowns", "interview me about this change", "brainstorm the approach for a task I am about to build", "help me scope this", "what am I missing here", "plan this before I build", "I know nothing about X in this codebase". When a territory profile is present it reads memory and the graph first and feeds the map into the build pipeline; without one it degrades to a plain general map. Not for trivial or mechanical tasks with clear acceptance criteria, single-fact lookups, pure debugging of a known symptom, or pure code review, those have their own tools. Do not use it to run the build itself; the map is the deliverable and implementation remains a separate task, whether it has not started or is already underway.
Also in 5 projects that use this
- +6 more from dzhng
Other versions43 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 43
- tldraw/tldraw★ 51ksame name
- tabler/tabler★ 42ksame name
- videojs/video.js★ 40ksame name
- PrefectHQ/prefect★ 24ksame name
- comet-ml/opik★ 23ksame name
- coder/coder★ 17ksame name
- browseros-ai/BrowserOS★ 14ksame name
- Agenta-AI/agenta★ 4.9ksame name
- liveblocks/liveblocks★ 4.8ksame name
- dzhng/jevgrep★ 3.2ksame name
33 more not listed
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository9 other files it gives its agents.
- AGENTS.mdAGENTS.md
- docs/AGENTS.mdAGENTS.md
- .claude/CLAUDE.mdCLAUDE.md
- .github/copilot-instructions.mdCopilot instructions
- create-migrationSkill
- write-docsSkill
- write-translationsSkill
- create-migrationSkill
- write-translationsSkill
The file
Show its rules and text
We split this skill into the rules it gives an agent
13 rules. 0 are common, 12 are distinctive, and 1 is very specific to this project.
0
Common
Most of the 42 edited versions we read keep these.
12
Distinctive
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 15
Read the repo contract, the guide nearest the docs tree, the renderer's component registry, and the checks for that content.
- Very specific: line 20
Use the applicable locale’s resolved message catalogs for visible UI labels.
- Distinctive: line 31
Keep the note proportional to the change.
- Distinctive: line 43
Run actionable code examples in the documented environment.
- Distinctive: line 48
If observed behavior and docs disagree, identify which is wrong.
- Distinctive: line 55
Use PLAYBOOKS.md to choose a shape.
- Distinctive: line 60
Avoid introductions that only announce the page's contents.
- Distinctive: line 61
Make the path easy to follow: Put conditions before dependent actions.
- Distinctive: line 74
Avoid marketing claims, “simply”/“easy”/“just” judgments, status chatter, and exclamation marks in ordinary docs prose.
- Distinctive: line 85
Use lists for parallel choices and numbered steps for sequences, regardless of item count.
- Distinctive: line 92
Update every full locale supported by the affected content tree in the same change, following write-translations.
- Distinctive: line 101
Read the rendered page as someone entering from search, without the author's prior context.
- Distinctive: line 106
Run the repository's applicable content, locale, link, asset, and build checks.
“Common” is measured: kept by at least 6 in 10 of the 42 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Write documentation people can use
Help the reader make progress from the situation they are in. A newcomer needs a guided first success; an experienced user needs the missing step; an administrator needs consequences and permissions; a developer needs exact inputs and outputs. Support these needs with connected pages, not one page that assumes every reader has the same knowledge.
Discover before drafting
Read the repo contract, the guide nearest the docs tree, the renderer's component registry, and
the checks for that content. Tale has two content trees: docs/AGENTS.md governs the EN/DE/FR
product guides, while the design-system authoring contract
governs the English component guides and their live demos. Use the contract for the tree you edit.
Find the existing owner of the topic and its neighboring pages before adding another explanation.
Use the applicable locale’s resolved message catalogs for visible UI labels; shared controls may
read @tale/ui or @tale/marketing-ui catalogs beneath the service’s overrides.
In the task's planning note outside the clone, record:
- Reader, goal, starting knowledge, and the page's job: tutorial, how-to, explanation, reference, troubleshooting, or navigation.
- Current gaps and the facts that need checking: role, prerequisites, route, inputs, defaults, limitations, result, and recovery.
- Existing content to reuse or link, worthwhile visuals/examples, and affected locales and links.
Keep the note proportional to the change. Reuse an existing task note; no separate notes skill is required. For a broad overhaul, map readers to their core tasks and record coverage and verification status. Page count and word count are not measures of completeness.
Establish the facts
Drive the real UI before describing a workflow. Use a disposable, seeded local organization when available. Follow the documented starting point through the result; inspect relevant empty, loading, permission, and failure states. Check keyboard access for the controls you describe. Source and tests clarify hidden defaults, limits, and persistence; they do not replace using the UI. A delegated observer may provide a concrete route/action/result record with the relevant version.
Run actionable code examples in the documented environment. Preserve the actual shape of results; sanitize secrets and replace unstable IDs consistently. Label excerpts, placeholders, optional steps, and setup dependencies. Do not invent outputs, timings, permissions, guarantees, or causes to make prose sound complete. Record any verification limitation explicitly.
If observed behavior and docs disagree, identify which is wrong. Fix an in-scope defect with a regression test, then rerun the flow. Otherwise record the unresolved finding and document the verified limitation. Respect the user's authorized scope; a docs task alone is not permission to change production data or send invitations.
Write for the task
Use PLAYBOOKS.md to choose a shape. The shape supports the work; it is not a mandatory section template.
- Open with useful information. State the outcome or the essential distinction directly. One sentence can be enough. Add audience, prerequisites, or context where they resolve uncertainty. Avoid introductions that only announce the page's contents.
- Make the path easy to follow. Put conditions before dependent actions. Explain unfamiliar terms where first needed. Keep the ordinary path visible; link deeper concepts and exhaustive reference. Place a warning before the action whose consequence matters.
- Add detail where a reader would hesitate. Explain what to enter, how to choose, what changes, when to wait, how to recognize completion, and how to recover. Use a believable worked example when an abstract explanation leaves the choice unclear.
- Use connected prose. Address the reader as a calm peer. Prefer active verbs and concrete nouns; vary sentence length naturally. Use imperatives for required actions and “you can” for a real optional capability. Name the purpose once; do not repeat “To…” before every click.
- Stop when the reader has what they need. A final result, a relevant next link, or the last reference entry can be the ending. Add a recap only when it helps retain a complex idea. Do not manufacture a closing heading or repeat the introduction.
Avoid marketing claims, “simply”/“easy”/“just” judgments, status chatter, and exclamation marks in ordinary docs prose. Preserve literal UI strings and quoted technical syntax. Teach the reason for an action where it changes a decision; obvious controls do not need a paragraph of justification.
Choose useful evidence and structure
Images should clarify a location, unfamiliar state, relationship, or result. Do not add a hero image or an image to every step by default. Keep the written instructions complete without the image. Read SCREENSHOTS.md before changing an asset; every shipped screenshot must be reproducible through the repository's capture pipeline.
Use lists for parallel choices and numbered steps for sequences, regardless of item count. Use components when they improve scanning or comparison; COMPONENTS.md covers their tradeoffs. Keep prerequisites and recovery instructions visible. Do not turn every useful detail into a callout. Read MECHANICS.md for links, examples, and Markdown conventions.
Localize the complete experience
Update every full locale supported by the affected content tree in the same change, following write-translations. Preserve the reader's task, factual content, examples, warnings, and navigation. Sentence construction and paragraph boundaries should read naturally in each language. Tale's heading/component parity is a delivery guard, not a requirement to translate sentence by sentence. Verify localized labels, links, alt text, and captions. An English-only guide does not acquire translated body routes merely because its shared chrome has multilingual catalogs.
Review and prove
Read the rendered page as someone entering from search, without the author's prior context. Complete the task from the instructions. Check whether a newcomer can locate the starting point, whether an experienced reader can find a specific answer, and whether all required facts are available without opening optional panels or interpreting an image.
Run the repository's applicable content, locale, link, asset, and build checks. Inspect important pages at narrow and wide widths, including images, tables, code, keyboard focus, and heading order. Review prose separately from mechanical checks: a green suite does not prove accuracy, flow, or native translation. Report what was observed and any remaining limits.
References
- PLAYBOOKS.md: choosing page purpose and the needed level of detail.
- COMPONENTS.md: component decisions and accessible structure.
- SCREENSHOTS.md: visual selection, capture, and review.
- EXAMPLES.md: edits that improve flow without adding filler.
- MECHANICS.md: source and formatting conventions.
This method draws on Diátaxis, Google's procedure guidance, and Microsoft's accessible writing guidance. The repository defines its implementation; these sources guide editorial judgment.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

