readme-writing
Use when the user asks to write, rewrite, restructure, review, or correct a project's README.
Skill by jhonDoe15
Chalk Labs
Need help deploying readme-writing?
Chalk Labs sets up agents like this inside real teams: connected to your tools, tested on your data, with a human in the loop where it matters.
Install
Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.
npx agentmarket install skill/jhondoe15-readme-writing
Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.
Get the weekly diff
The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.
About
Use cases: writing
Reviews
Write a reviewNo reviews yet. Installed it? Reviews come from verified installers only. How reviews work
Permissions
What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.
- Tools
- No tools found
- Secrets
- None asked for
- Install scripts
- None found
- Runs as
- Python package
Network hosts it talks to 1
codeload.github.com· found in install command
Files it touches
~/.claude/skills/readme-writing
What's inside
- Skill name
readme-writing- Description
- Use when the user asks to write, rewrite, restructure, review, or correct a project's README.
- License
- MIT
- Instructions
- 151 lines in SKILL.md · view on GitHub
Before installing
- Ships executable scripts
Files (27)
- SKILL.md
- SOURCES.md
- colprune.js · evals/fixtures/colprune/bin/
- contacts.csv · evals/fixtures/colprune/examples/
- package.json · evals/fixtures/colprune/
- README.md · evals/fixtures/snag/
- snag.js · evals/fixtures/snag/bin/
- philosophy.md · evals/fixtures/snag/docs/
- app.py · evals/fixtures/snag/examples/sample/src/
- web.ts · evals/fixtures/snag/examples/sample/src/
- package.json · evals/fixtures/snag/
- debug-call.js · evals/fixtures/snag/src/checks/
- duplicate-block.js · evals/fixtures/snag/src/checks/
- index.js · evals/fixtures/snag/src/
- tamemark.js · evals/fixtures/tamemark/bin/
- tamemark.browser.js · evals/fixtures/tamemark/dist/
- design.md · evals/fixtures/tamemark/docs/
- input.md · evals/fixtures/tamemark/examples/
- run.js · evals/fixtures/tamemark/examples/
- package.json · evals/fixtures/tamemark/
- build-browser.js · evals/fixtures/tamemark/scripts/
- core.js · evals/fixtures/tamemark/src/
- footnotes.js · evals/fixtures/tamemark/src/
- index.js · evals/fixtures/tamemark/src/
- outcome.json · evals/
- role.json · evals/
- trigger.json · evals/
SKILL.md (first lines)
# README Writing A README is the page a stranger reads to decide whether a project is for them, and then to start using it. This Skill owns the README's substance: what goes in it, in what order, and whether each claim is true. The `to-humans` Audience outcome owns how its sentences read; the managing agent selects it independently, and this Skill does not invoke it. ## Routing Select `readme-writing` when the deliverable is a project README (`README.md` or its equivalent at the root of a repository, package, or workspace member). Direct invocation of `readme-writing` wins for its README deliverable. Select another outcome for: - agent-facing guidance (`AGENTS.md`, `CLAUDE.md`): `agents-file-writer`; - reference documentation pages, changelogs, release notes, and contributor guides. No suite outcome owns these; handle them as ordinary requested work. Before choosing a branch, name the target file. When more than one README could be meant (a monorepo package, a profile or organization README), ask which one. ## Choose the branch - **Create or restructure:** the user asks for a new README, a rewrite, or a new structure. Run steps 1 to 5. - **Targeted correction:** the user asks to change named content (fix the install command, add a section, update a claim) without asking for a new structure. Edit only that content, then run step 5 for the claims you changed. - **Review:** the user asks to assess the README, or says not to change it. Run steps 1 to 3 and step 5 checks 1 to 4 without changing any file, then return findings ranked by impact on a stranger, each quoting the README location and citing the source file, run, or reference that shows the problem. ## 1. Establish the project's purpose Write down in one sentence what the project is for and what makes it different, in the maintainers' own terms. Take it from their stated goals, design or philosophy docs, and the code's behavior, not from the current README. List the facts a reader needs to adopt it: what it supports, what it requires, how to install and run or use it, its maturity, and its license. Trace each fact to a source file. If no maintainer statement of purpose exists, propose one from the code and docs, mark it as proposed, and ask the user to confirm it before writing. If the maintainers' stated purpose and the current README disagree, report both and ask which governs. Complete when the purpose sentence is confirmed or sourced and every fact has a source. ## 2. Read reference READMEs Read the READMEs of five or more widely used projects of the same kind (for a linter, other linters and formatters; for a library, comparable libraries; for an application, comparable applications). Fetch the raw files; do not rely on memory. Fewer than five is fine when the kind has fewer well-known projects; say so. For each, note how it opens, what proves its claim, what it leaves to its docs, and how it handles anything the project has in common with it (variants, maturity, a strong stance). Where the references agree with a default in step 3, keep the default. Where most of them depart from it in the same way, follow them and record why. Complete when the references read and any departures from step 3's defaults are written down. ## 3. Choose content and order Default order, top first. Items marked "when" apply only when their condition holds. 1. **Name and identity line.** One sentence naming a category the reader already knows plus what sets this project apart. The purpose leads, not a feature. 2. **The problem.** One short paragraph a stranger can follow. When the purpose sentence names something the user gives up or must accept (strict rules, fixed formatting, fewer options), state that trade here, with what the user gets for it. 3. **Proof.** One concrete demonstration right under the pitch, taken from a real run of the current code: a command and its output for a tool, a short usage snippet and its result for a library, a screenshot for an application, or a benchmark. Never invent or edit output beyond trimming whole lines. Show any file path or setup the example depends on. 4. **Highlights.** A short list of one-line reasons to choose the project. Each must add something the identity line and the proof do not already show. 5. **Getting started.** Requirements, then the fewest commands that get a first result. When there are variants (languages, platforms, frameworks, modes), one compact table. Then the first integration step most users need (for example CI, an editor, or a framework plugin). 6. **Objections**, when the project enforces choices users may resist: answer "what if I disagree with X?" in the project's own terms. 7. **Documentation links**, each with a few words on what it covers. 8. **Status**, when the version is below 1.0 or the project calls itself experimental: what may change, and what to pin. 9. **Contributing, License, Acknowledgements.** Keep what the existing README carries for its community (badges, translation links, sponsor and contributor sections, a code of conduct link) unless the user asks to remove it. Move to the docs and link instead: command-line flag lists, exit codes, configuration keys, versions of bundled third-party tools, internal paths, architecture, project history, issue or pull request numbers, and anything only one variant's users need. Keep one of these in the README only when most readers need it to get a first result, and say why. Complete when every fact from step 1 has a place in the README or a linked doc, and every README section serves the stranger's decide-then-start path. ## 4. Write Write the README in that order. Keep each claim true for every variant it covers: when it holds for some variants only, scope it in the sentence or move it to the variant's doc. When another file repeats README text (a package description, keywords, a doc that says it mirrors the README), list those files and update them in the same change. Create a linked doc only when the user asked for the README work and the moved material has no existing home; name the new path. Complete when the README and the listed files are written. ## 5. Verify 1. Check every claim against the code and docs, row by row for tables and lists of variants. 2. Run the commands the README tells a reader to run, in a scratch checkout or sandbox. Do not run remote install scripts or commands that change the user's machine or accounts without authorization; report those as unverified. 3. Confirm the proof matches a real run, including line numbers against the code shown. 4. Resolve every link and anchor. 5. Get two independent adversarial reviews of the draft, one for factual accuracy against the code
From jhonDoe15/skills, MIT.
More from jhonDoe15/skills
- 0.570score
model-routing
jhonDoe15
Use before the first search or edit of work that needs more than two steps, and whenever bounded work might be handed to another model through `pi-dispatch`.
SkillAFreshness: AClaude Code +9 - 0.564score
incident-investigation
jhonDoe15
Evidence-led, read-only investigation of production incidents and hard-to-localize system or application bugs.
SkillAFreshness: AClaude Code +9 - 0.561score
skill-writing
jhonDoe15
Agent Skill outcome for creating or revising Agent Skills; excludes general agent-facing artifacts, fresh-context handoffs, installation, publication, and…
SkillAFreshness: AClaude Code +9 - 0.529score
engineering-guidance
jhonDoe15
Engineering assessment of an existing technical approach or change for maintainability and standards; also used by declared design, implementation, and review…
SkillBFreshness: BClaude Code +9 - 0.526score
implementation-planning
jhonDoe15
Plan one bounded implementation ticket before code: produce a reviewed technical design and executable consumer-behavior TDD plan for another agent.
SkillBFreshness: BClaude Code +9 - 0.526score
architecture-review
jhonDoe15
Architecture review of existing code, a supplied technical design, or a PR change range.
SkillBFreshness: BClaude Code +9 - 0.524score
code-review
jhonDoe15
Reviews a complete Ticket outcome through fresh Domain and Engineering/Design lenses, then returns an evidence-backed read-only Review brief.
SkillBFreshness: BClaude Code +9 - 0.524score
skill-evaluation
jhonDoe15
Private evaluation dependency for Agent Skill behavior claims.
SkillBFreshness: BClaude Code +9 - 0.524score
implement
jhonDoe15
Use when a ready implementation blueprint and bounded scope need one isolated, TDD-driven patch with validation and a correction-ready implementation handoff.
SkillBFreshness: BClaude Code +9 - 0.524score
pr-carver
jhonDoe15
Use when assessing an existing Git branch or pull request and recommending one PR, parallel PRs, or an ordered stack from its prerequisites, migration…
SkillBFreshness: BClaude Code +9 - 0.523score
slice-plan
jhonDoe15
Private set-level decomposition consumed by Carve; not a user-facing planning outcome.
SkillBFreshness: BClaude Code +9 - 0.523score
maintaining-agent-guidance
jhonDoe15
Use when finishing or merging a substantial project change that may require durable updates to agent guidance.
SkillBFreshness: BClaude Code +9
Signals
- Stars
- 0
- Forks
- 0
- Open issues
- 16
- Last commit (this folder)
- Oct 13, 2026 (16 days ago)
From GitHub (fetched Oct 24, 2026).
Details
- License
- MIT
- Repository
- github.com/jhonDoe15/skills
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
- Source
- GitHubgithub-graphql
More from jhonDoe15
- 0.570score
model-routing
jhonDoe15
Use before the first search or edit of work that needs more than two steps, and whenever bounded work might be handed to another model through `pi-dispatch`.
SkillAFreshness: AClaude Code +9 - 0.564score
incident-investigation
jhonDoe15
Evidence-led, read-only investigation of production incidents and hard-to-localize system or application bugs.
SkillAFreshness: AClaude Code +9 - 0.561score
skill-writing
jhonDoe15
Agent Skill outcome for creating or revising Agent Skills; excludes general agent-facing artifacts, fresh-context handoffs, installation, publication, and…
SkillAFreshness: AClaude Code +9 - 0.529score
engineering-guidance
jhonDoe15
Engineering assessment of an existing technical approach or change for maintainability and standards; also used by declared design, implementation, and review…
SkillBFreshness: BClaude Code +9 - 0.526score
architecture-review
jhonDoe15
Architecture review of existing code, a supplied technical design, or a PR change range.
SkillBFreshness: BClaude Code +9 - 0.526score
implementation-planning
jhonDoe15
Plan one bounded implementation ticket before code: produce a reviewed technical design and executable consumer-behavior TDD plan for another agent.
SkillBFreshness: BClaude Code +9
Done for you
Rather not set it up yourself?
Chalk Labs scopes, builds and runs agents for your team: the task audit, the eval harness, the integrations and the weekly ops report. Fixed scope, priced before we start.
See how we build AI agents →Read the case studies
Chalk Labs is a two-founder AI and growth studio. We build AI agents, CRMs and MVPs for Web3 and AI companies, and have shipped 50+ product and token launches.
- launches shipped
- 50+
- launches shipped
- media placements
- 500+
- media placements
- founders on every account
- 2
- founders on every account
Press we've placed: Forbes · CoinDesk · Decrypt · Cointelegraph · The Block
Rahil Jain, Co-founder, Chalk Labs. Ex-Goldman Sachs engineer. You talk to the person who builds it.
Get readme-writing running in your business
Tell us what you want to automate. A founder replies within one working day with a plan, not a pitch deck.
Cite this page
readme-writing skill. Agent Mart. https://agentmart.dev/skills/jhondoe15-readme-writing (last updated 2026-10-24)