technical-writing
Skill454 starsChanged 9 days agoMITView on GitHub
Write, organize, name, edit, or review Enterprise developer documentation, specifications, and technical instructions; ground claims in current source and finish with a plain-language cleanup.
142 other projects use a version of this.
Adopted by
Most recently taken up by
- Yeachan-Heo/oh-my-claude-sisyphustheir own version, first seen Nov 8, 2026
- Loner1024/skillstheir own version, first seen Nov 8, 2026
- lklimek/claudiustheir own version, first seen Nov 8, 2026
- justinyoo/skillstheir own version, first seen Nov 8, 2026
- thrillmade/agent-skillstheir own version, first seen Nov 8, 2026
- metraton/gaiatheir own version, first seen Nov 8, 2026
Used with
Projects with it also use ketchup-video, principle-explain-the-number 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.
tristanphvnSkillketchup-video
Catch the user up like /keel:ketchup, with a short silent video of slides for what happened and the waiting calls as text under it. Use for /keel:ketchup-video.
Also in 14 projects that use this
ianclearySkillprinciple-explain-the-number
Apply before you trust, report, or act on a number you measured: a speedup, a regression, a throughput, a latency, or an eval result. Find what limits it, and rule out that it measured something other than the work you think.
Also in 29 projects that use this
eryzerzSkillmake-bot-ui
Use when building a custom UI (page, dashboard, buttons) that should wake a Grok Bot over a webhook, when the user must provide a webhook sender key, or when exposing that UI on Tailscale.
Also in 20 projects that use this
jaydubya818Skillprinciple-attack-the-premise
Apply when repeated fixes fail and a shared assumption may be wrong.
Also in 35 projects that use this
jaydubya818Skillprinciple-test-behavior-not-implementation
Apply when tests mirror internals instead of asserting observable outcomes.
Also in 36 projects that use this
ianclearySkillpoteto-help
Guides users through pstack setup, /poteto-mode, and picking the skill, playbook, or principle for a task. Type /poteto-help with a question.
Also in 14 projects that use this
- +5 more from jaydubya818
Other versions142 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 142
- Yeachan-Heo/oh-my-claude-sisyphus★ 41ksame name
- Yeachan-Heo/oh-my-claudecode★ 41ksame name
- vercel/eve★ 5.8ksame name
- backnotprop/pstack★ 3.8ksame name
- chattocorp/chatto★ 2.9ksame name
- cashew-labs/libretto★ 898same name
- saffron-health/libretto★ 883same name
- moeru-ai/xsai★ 667same name
- proffesor-for-testing/agentic-qe★ 524same name
- NPC-Worldwide/npcsh★ 502same name
132 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 repository13 other files it gives its agents.
- AGENTS.mdAGENTS.md
- autoreviewSkill
- console-uiSkill
- design-reviewSkill
- deslopSkill
- enterprise-testingSkill
- local-devSkill
- mermaid-diagramsSkill
- oce-cookbookSkill
- oceintegSkill
- pr-walkthroughSkill
- specSkill
- test-auditSkill
The file
Show its rules and text
We split this skill into the rules it gives an agent
26 rules. 0 are common, 20 are distinctive, and 6 are very specific to this project.
0
Common
Most of the 134 edited versions we read keep these.
20
Distinctive
This author's own choices. Worth a look.
6
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 9
Use for repository documentation, including READMEs, guides, references, flow docs, specifications, and technical PR descriptions.
- Distinctive: line 22
Verify claims, examples, defaults, paths, and failure behavior.
- Distinctive: line 24
Before handing off any writing, complete the required plain-language pass.
- Distinctive: line 30
Keep official resource names and public identifiers.
- Distinctive: line 33
Use present tense, active voice, concrete nouns, and one term per concept.
- Distinctive: line 39
Delete repeated background, meta-commentary, and generic benefits.
- Very specific: line 40
Treat all, only, and never as literal claims.
- Very specific: line 41
Use must for requirements and can for options.
- Distinctive: line 42
Use sentence-case, action-specific article headings and descriptive links.
- Distinctive: line 52
Run this pass whenever you write, edit, or review prose, including small changes, PR descriptions, and review findings.
- Distinctive: line 61
Delete filler, canned transitions, empty claims, and repeated explanations.
- Distinctive: line 62
Remove contrasts or caveats the reader does not need.
- Distinctive: line 63
Put the main point first. Split a sentence when the reader has to untangle several actions, actors, or conditions.
- Distinctive: line 65
Check the edit against the original and the evidence.
- Distinctive: line 71
When adding, grouping, moving, or renaming documentation, read Organization and naming.
- Distinctive: line 79
Keep quickstarts and main guides focused on the normal supported workflow.
- Very specific: line 104
Use YAML frontmatter for page metadata.
- Very specific: line 106
Write tags as a list of strings.
- Very specific: line 110
Keep verification status in the tag and record the human review date in last_verified (YYYY-MM-DD).
- Very specific: line 118
Use the Markdown frontmatter field kind to select a template from the documentation taxonomy.
- Distinctive: line 132
Keep shared setup in its owning guide and preserve existing heading anchors.
- Distinctive: line 138
Keep the template's contents list below the title and update its links with the page.
- Distinctive: line 139
Put implementation-specific setup and settings in the linked implementation page.
- Distinctive: line 140
When adding a template, add its kind, selection criteria, and asset link to this table and the documentation taxonomy in the same change.
- Distinctive: line 151
Put permissions, secret handling, destructive effects, concurrency limits, timeouts, ordering, retries, and recovery beside the affected …
- Distinctive: line 167
Check commands, examples, terminology, local links, and navigation.
“Common” is measured: kept by at least 6 in 10 of the 134 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
Technical writing
Use for repository documentation, including READMEs, guides, references, flow
docs, specifications, and technical PR descriptions. Read the applicable
AGENTS.md first; it owns documentation destinations, length limits, and
repository terminology. This skill needs no global tools or skill installation.
Establish the reader and evidence
- Identify the reader, intended action or decision, page scope, and source of truth. Read current code, schemas, tests, command output, and the relevant diff.
- Choose the requested mode: create from evidence, edit while preserving useful facts and the author's intent, or review with findings before proposed fixes.
- Lead with what the reader can accomplish and one recommended path. Mention implementation first when explaining architecture or internals is the task.
- Verify claims, examples, defaults, paths, and failure behavior. Label material uncertainty and name the missing evidence; polished prose is not verification.
- Before handing off any writing, complete the required plain-language pass.
Write precise prose
Apply ASD-STE100 (Simplified Technical English). Before and after a documentation update, complete the terminology and limits pass. Keep official resource names and public identifiers. Check each replacement against its meaning and current source.
- Use present tense, active voice, concrete nouns, and one term per concept. Expand unfamiliar abbreviations at first use; use established product names.
- Define unfamiliar concepts through their owner, action, and observable role. Replace “the controller handles deployment” with what it reads, decides, and starts, when those details matter to the reader.
- Make every sentence help the reader decide, act, or understand a boundary. Delete repeated background, meta-commentary, and generic benefits.
- Treat
all,only, andneveras literal claims. Name the surface they cover. Usemustfor requirements andcanfor options; preserve consequential limits. - Use sentence-case, action-specific article headings and descriptive links. Explain why when it changes a decision or prevents a failure.
- Give each contract, default, and behavior one owning page. Link its definition from other pages rather than copying it into competing references.
- Make diagrams agree with prose: actors and resources are nodes, containment uses regions, and arrows describe interactions. Do not imply deployment proof from source alone or rely on color alone to convey meaning.
Always deslop the writing
Run this pass whenever you write, edit, or review prose, including small changes, PR descriptions, and review findings. The user does not need to request it. For edits, clean up the prose in scope and read the surrounding text for flow. For a review-only task, clean up your feedback and flag wording in the document when it obscures meaning; do not silently rewrite the source.
- Name who does what and under which conditions. Replace vague verbs and noun piles such as “performs readiness observation” with “checks whether the workload is ready.” Prefer familiar words; define necessary technical terms.
- Delete filler, canned transitions, empty claims, and repeated explanations. Remove contrasts or caveats the reader does not need. Avoid invented labels.
- Put the main point first. Split a sentence when the reader has to untangle several actions, actors, or conditions; give each paragraph one main point.
- Check the edit against the original and the evidence. Preserve scope, required actions, permissions, warnings, failure behavior, and uncertainty. Do not make an unsupported claim sound certain or cut a useful detail merely to shorten it.
Organize and name development docs
When adding, grouping, moving, or renaming documentation, read Organization and naming. Choose one home based on the reader's task. Use short sidebar labels and give the article enough context to make sense when opened directly. Check repository guidance and the current navigation before treating a proposed layout as implemented.
Choose the smallest useful page
Keep quickstarts and main guides focused on the normal supported workflow. Put platform-specific failures, uncommon environment workarounds, and extended diagnostic steps in the owning troubleshooting page or section. Link to that guidance with a short, symptom-based pointer beside the affected step; do not front-load the main guide with edge cases. Keep required prerequisites, security boundaries, destructive effects, and recovery needed for the normal workflow beside the action they affect.
| Page | Include |
| Overview or README | Reader outcome, scope, recommended starting path, and links to detail. |
| Quickstart | Prerequisites, minimum configuration, one runnable example, expected result, and next step. |
| Operator guide | Inputs, command, completion evidence, and recovery in execution order. Keep required inputs separate from optional inputs and show defaults beside options. |
| API or CLI reference | Purpose, permissions, exact inputs, defaults, constraints, outputs, side effects, errors, and examples. Keep generated schemas owned by their generator. |
| Testing guide | Setup, fixtures and permissions, success/failure proof, cleanup, and differences from production. |
| Troubleshooting | Observable symptom, first discriminating check, likely causes, concrete fix, and proof of recovery. |
| Architecture or specification | Selected model, owners, boundaries, invariants, tradeoffs, and proof; read specification guidance. |
| Runtime flow | Trigger, source pointers, runtime order, state and ownership transitions, decisions, failures, and handoff; follow the repository's local-dev flow contract. |
Omit empty or irrelevant sections. Split independently useful topics when a page mixes too many reader tasks. Keep the root documentation map and affected links current, following repository page ownership and length limits.
Frontmatter and tags
Use YAML frontmatter for page metadata. Read Frontmatter and tags
for supported fields, values, and their owners before adding or changing metadata.
Write tags as a list of strings. The nolimit tag exempts its page from both
word-count thresholds while retaining count reporting; apply it only with the
human approval required by AGENTS.md. Add human_verified to the tags only
when a human has verified the document, for example tags: [human_verified].
Keep verification status in the tag and record the human review date in
last_verified (YYYY-MM-DD). Update that date only after another human review.
Preserve existing tags and metadata.
Change generated frontmatter through its generator. Snippet tags do not exempt
pages that include the snippet.
Templates for specific document types
Use the Markdown frontmatter field kind to select a template from the
documentation taxonomy. Copy the template for a new
page. When updating an existing page, preserve its links and heading anchors.
| Kind | Use it for | Template |
reference/ci | One test workflow's triggers, scenarios, execution, and result limits | CI check Markdown |
reference/driver | The shared contract for one Driver capability | Driver contract Markdown |
reference/driver/impl | One Driver implementation's provisioning, configuration, and operating limits | Driver implementation Markdown |
For a test workflow detail page, use kind: reference/ci and keep the template's six
sections in order: Overview, Trigger, Scenarios, Execution, Results and limits,
and Files. Follow the CI authoring guidance
for scenario descriptions, descriptive names, and exact GitHub job mappings.
Keep shared setup in its owning guide and preserve existing heading anchors.
Publication-only workflows, preview builds, cache maintenance, and security
scanning belong in their owning contributor or release guides, not this template.
For a Driver contract, keep the template's sections in order and use the Driver authoring guidance to check coverage. Keep the template's contents list below the title and update its links with the page. Put implementation-specific setup and settings in the linked implementation page. When adding a template, add its kind, selection criteria, and asset link to this table and the documentation taxonomy in the same change.
Make examples usable
Show realistic, safe inputs and exact identifier types. Mark placeholders clearly, quote YAML values when needed, and specify each code block's language. Include the working directory, prerequisites, invocation, and expected success output when they are necessary to run a command. Verify examples when feasible; report when they have not been executed.
Put permissions, secret handling, destructive effects, concurrency limits, timeouts, ordering, retries, and recovery beside the affected step when mistakes have consequences. Separate development, test, and production behavior. Never include real credentials. Remove explanation before removing information needed for a command to succeed safely. Keep internal orchestration in its owning flow or reference unless the operator needs it to make a decision.
Edit and verify
Correct inaccurate or unsafe claims first, add missing requirements or failure handling, then remove repetition and tighten prose. Update current docs with the behavior they describe; mark a page stale with a source-of-truth link if a full update cannot be completed. Preserve historical specs and user-owned Manual Notes in source. The site omits document Changelogs and empty Manual Notes. Do not rewrite history to match later implementation.
Check commands, examples, terminology, local links, and navigation. For documentation-only changes, use formatting, builds, link checks, and visual inspection; do not add or run tests. Report checks run and gaps.
For a review, cite the conflicting text, explain its consequence, and suggest the smallest correction. Order findings by reader impact, distinguish correctness from optional polish, and state what evidence was checked. Say when no actionable findings remain.
Provenance
Adapted from Docy's core technical-writing and document-lifecycle guidance, developer documentation, concise instructions, and specification references. The repository maintains this self-contained selection; see the developer-skills catalog for source details. Docy's CLI and personal installation are not required.
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.

