agentleFS
unclejobs-aiunclejobs-ai/second-claude-code/skills/write/SKILL.md

write

Skill41 starsChanged 7 days agoMITView on GitHub

Use when producing newsletters, articles, reports, shorts, or social content

35 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use read, hunt and 7 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.
  • MicrosoftSkill

    read

    Read pages from a document. Returns MinerU markdown (preferred) or PyPDF text, with optional page screenshots and sub-image fetching. When markdown is available, a figure_images_meta catalog is included — use the figures parameter to fetch specific sub-image pixels. Triggers include "read page 5", "what does page 12 say", "show me the chart on page 3", "extract pages 3-5".

    Also in 6 projects that use this

  • ninehillsSkill

    hunt

    Finds root cause before applying fixes for errors, crashes, regressions, failing tests, broken behavior, and screenshot-reported defects. Use when users report in any language errors, crashes, broken behavior, regressions, failing tests, screenshot evidence, or something that used to work and now fails. Not for code review or new features.

    Also in 4 projects that use this

  • vltingSkill

    think

    Lightweight plan-to-execution pipeline. Explores codebase, runs council deliberation, presents plan for approval, then executes via /run.

    Also in 7 projects that use this

  • genomewalkerSkill

    health

    Soul system health check with remediation

    Also in 5 projects that use this

  • kt057Skill

    check

    lint, typecheck, test を一括実行して結果をサマリーする

    Also in 5 projects that use this

  • genomewalkerSkill

    learn

    Learn from web sources, generate SKILL.md, and store key patterns in soul memory. Trigger keywords: learn, skill-gen, web-learn, documentation, tutorial.

    Also in 8 projects that use this

The file

Show its rules and text

We split this skill into the rules it gives an agent

14 rules. 0 are common, 8 are distinctive, and 6 are very specific to this project.

  • 0

    Common

    Most of the 31 edited versions we read keep these.

  • 8

    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
  1. Very specific:

    Run /scc:review --preset quick UNLESS --skip-review flag is set.

    line 34
  2. Very specific:

    When --input is provided (or when called from a pipeline with input_from), the file content is injected as source context and …

    line 52
  3. Very specific:

    If .data/soul/SOUL.md exists, read its ## Tone Rules section and treat those rules as non-negotiable constraints on top of the selected …

    line 62
  4. Distinctive:

    When user-specified length conflicts with format minimums:

    line 90
  5. Distinctive:

    If user insists on short article, respect user intent but note the override in output metadata.

    line 93
  6. Distinctive:

    Never silently truncate or silently exceed the user's request.

    line 95
  7. Very specific:

    After the review step (or after drafting if --skip-review), save the final content:

    line 99
  8. Distinctive:

    Write the full content using the Write tool.

    line 103
  9. Very specific:

    When --publish notion is set: requires a configured Notion MCP connection (mcp__claude_ai_Notion__notion-create-pages).

    line 106
  10. Distinctive:

    Do not skip research unless real sources are already supplied.

    line 110
  11. Distinctive:

    Do not start writing before loading the format spec file.

    line 111
  12. Very specific:

    Do not output content without running the review step (unless --skip-review is explicitly set).

    line 112
  13. Distinctive:

    Do not miss the CTA in shorts or conclusions.

    line 113
  14. Distinctive:

    Do not ship the reviewed version without addressing Critical and Major issues.

    line 114

“Common” is measured: kept by at least 6 in 10 of the 31 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.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Iron Law

Never publish content without research and review.

Red Flags

  • "I'll skip the research step, I know enough about this topic" → STOP, because writing without sources falls back to training knowledge silently — warn the user or provide --input.
  • "The review found issues but they're minor, ship it" → STOP, because Critical and Major findings must be addressed before output — review is mandatory, not advisory.
  • "I'll figure out the format as I go" → STOP, because the format spec file must be loaded before drafting — writing without it produces structure drift.
  • "This is only 800 words but it covers everything" → STOP, because format minimums (newsletter 10000, article 4000, report 5000 chars of body) are hard rules — negotiate length with the user, never silently truncate.
  • "The voice doesn't matter for this piece" → STOP, because tone inconsistency is the top reader-trust killer — load the voice guide and verify the voice checklist before finalizing.

Write

Produce content with automatic research and review unless the caller explicitly skips those steps.

When to Use

  • The user asks for a draft, article, newsletter, report, short-form script, or social post
  • Another skill needs polished written output

Workflow

  1. Run /scc:research UNLESS --skip-research flag is set OR source material is already provided via --input. Guard: If --skip-research is set but no --input is provided and no source material exists in the conversation, warn the user: "No source material provided. Output will rely on training knowledge only. Continue?" Do NOT silently fall back to training knowledge.
  2. Load format spec AND voice guide: read references/formats/{format}.md (required — abort if missing) and references/voice-guides/{voice}.md (recommended). If the voice guide file doesn't exist, proceed using the default voice characteristics from the Voices table mapping below.
  3. Draft content following loaded format spec and voice constraints.
  4. Run /scc:review --preset quick UNLESS --skip-review flag is set. This step is MANDATORY by default.
  5. Address all Critical and Major review findings. Re-read the review output and apply fixes.

Options

FlagValuesDefault
--formatnewsletter|article|shorts|report|social|card-newsarticle
--voicepeer-mentor|expert|casualformat-specific
--publishnotion|filefile
--skip-researchflagoff
--skip-reviewflagoff
--langko|enko
--inputfile pathnone
--constraintscomma-separated stringsnone

--constraints: Review findings injected as hard requirements for this draft. Each entry is enforced as a non-negotiable rule during drafting and is not open to negotiation with the user. Passed automatically by the PDCA Act→Do route; can also be supplied manually.

When --input is provided (or when called from a pipeline with input_from), the file content is injected as source context and --skip-research is implied.

Soul-Aware Voice Resolution

Voice is resolved in this priority order:

  1. --voice flag (explicit override — always wins)
  2. ## Tone Rules in .data/soul/SOUL.md (personal preferences — merge with voice guide)
  3. .data/soul/templates/STYLE.md defaults (bilingual baseline)
  4. Format voice guide from the Voices table below

If .data/soul/SOUL.md exists, read its ## Tone Rules section and treat those rules as non-negotiable constraints on top of the selected voice guide. Soul tone rules override format defaults but never override an explicit --voice flag.

Voices

VoiceDefault For
peer-mentornewsletter
expertreport, article
casualshorts, social, card-news

Voice Checklist

Before finalizing, verify:

  • At least one first-person example (peer-mentor) OR authoritative citation (expert) OR conversational aside (casual)
  • Tone is consistent throughout — no voice mixing
  • CTA present for shorts/social formats

Format Rules

  • newsletter: 6-stage arc, minimum 10000 chars of body, at least 2 research data points
  • article: minimum 4000 chars of body
  • report: minimum 5000 chars of body with numbered recommendations
  • shorts: around 1800 chars with a mandatory CTA
  • social: platform-optimized short post
  • card-news: slide-by-slide narrative with visual direction

Length Negotiation

When user-specified length conflicts with format minimums:

  1. Inform the user: "Article format minimum is 4000 chars, you requested ~800."
  2. Offer alternatives: switch to shorts or social format, OR keep article at full length.
  3. If user insists on short article, respect user intent but note the override in output metadata.

Never silently truncate or silently exceed the user's request.

Auto-Save

After the review step (or after drafting if --skip-review), save the final content:

  • Path: .captures/write-{format}-{slug}-{YYYY-MM-DD}.md
  • {slug} = topic lowercased, spaces to hyphens, max 40 chars
  • Write the full content using the Write tool. Do NOT skip this step.
  • Tell the user the saved path.

When --publish notion is set: requires a configured Notion MCP connection (mcp__claude_ai_Notion__notion-create-pages). Create a Notion page with the content, then save the local .md file. If Notion MCP is not available, warn the user and save to file only.

Gotchas

  • Do not skip research unless real sources are already supplied.
  • Do not start writing before loading the format spec file.
  • Do not output content without running the review step (unless --skip-review is explicitly set).
  • Do not miss the CTA in shorts or conclusions.
  • Do not ship the reviewed version without addressing Critical and Major issues. Minor issues may be deferred.
  • Severity levels follow the shared taxonomy: Critical (ship-blocking), Major (significant gap), Minor (polish). Same as /scc:review.

Subagents

writer: { model: opus, constraint: "follow format rules, voice, and length minimums" }
editor: { model: opus, constraint: "address all Critical and Major review items" }

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.