Agent skill

Write AI Doc

by samuelhe52 in samuelhe52/AniShelf

Create and maintain curated AniShelf docs-ai/ records for substantial features and non-trivial, decision-shaping fixes (numbered entries with 000-plan.md before implementation and 001-action.md…

Apache-2.0Auto-check passedMobile

Install Write AI Doc

skills CLI
$ npx skills add samuelhe52/AniShelf --skill write-ai-doc -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install samuelhe52/AniShelf write-ai-doc --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/samuelhe52/AniShelf.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.agents/skills/write-ai-doc .claude/skills/write-ai-doc && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
write-ai-doc
GitHub stars
140
Token cost
~1.8k tokens
SKILL.md length
715 words
Files
1
Skills in repo
2
Repo updated
First seen
Licence
Apache-2.0

At a glance

Create and maintain curated AniShelf docs-ai/ records for substantial features and non-trivial, decision-shaping fixes (numbered entries with 000-plan.md before implementation and 001-action.md…

  • Works in 3 steps: New entry — plan first, before coding → Follow-up on an existing entry (in-frame… → Large pivot / redesign
  • Routine investigations
  • SKILL.md covers When to write one, Workflow, Templates and Writing rules
  • Instructions only: no scripts, shell commands, URLs or credentials in SKILL.md

What it does

Write AI Doc is an agent skill from samuelhe52/AniShelf. Create and maintain curated AniShelf docs-ai/ records for substantial features and non-trivial, decision-shaping fixes (numbered entries with 000-plan.md before implementation and 001-action.md after). Do not use for reviews, audits, routine investigations, working notes, or status reports unless the maintainer explicitly asks for a docs-ai/ record.

Its SKILL.md is about 1.8k tokens, which your agent loads only when the skill is triggered. It is a single SKILL.md file with no bundled scripts.

It sits in Mobile, covering iOS development. It works with Xcode, SwiftUI and iOS. The repository describes itself as: A digital bookshelf for your anime – track, organize, and revisit your favorite series with ease. The licence is Apache-2.0.

When your agent uses it

  • Routine investigations
  • Status reports unless the maintainer explicitly asks for a docs-ai/ record

Example prompts

  • “/write-ai-doc”

Workflow steps

3 steps, taken from the step headings in SKILL.md.

  1. New entry — plan first, before coding
  2. Follow-up on an existing entry (in-frame fix or extension)
  3. Large pivot / redesign

What it can do on your machine

Read from SKILL.md and the folder at commit 5957f72. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    No scripts in the folder and no shell commands in SKILL.md (its code samples are markdown).

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    Links to these hosts (documentation or services it may open):

    • github.com

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Write AI Doc loads about 1.8k tokens when it runs. Until then it costs about 91 tokens; SKILL.md has 715 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~91
When it runs · the whole SKILL.md, loaded when a task matches
~1.8k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from samuelhe52/AniShelf at commit 5957f72, republished under its Apache-2.0 licence (© samuelhe52). 715 words, ~1,750 tokens.

Download SKILL.mdSave it as .claude/skills/write-ai-doc/SKILL.md (or your agent's skills folder).
name
write-ai-doc
description
Create and maintain curated AniShelf docs-ai/ records for substantial features and non-trivial, decision-shaping fixes (numbered entries with 000-plan.md before implementation and 001-action.md after). Do not use for reviews, audits, routine investigations, working notes, or status reports unless the maintainer explicitly asks for a docs-ai/ record.

Write AI Doc

docs-ai/ is AniShelf's curated product and design record, not a working-note log. A numbered folder is reserved for a substantial feature or a non-trivial, decision-shaping fix, each holding an RFC-like plan and an action log. Future humans and agents use it to answer "why is it built this way?", so entries must be selected deliberately and stay accurate against the code. Read docs-ai/README.md for the index before starting.

Adapted from the write-ai-doc skill in onevcat/Prowl. This copy is project-owned: edit it in place, and do not reinstall it from upstream.

When to write one

Create a new entry only when the work is either:

  • a substantial feature with an enduring product or architecture decision (for example, a new UI surface, a sync or persistence mechanism, or a schema change);
  • a non-trivial fix whose root cause, design decision, or resulting behavior must guide future implementation.

Do not create an entry merely because the work is detailed, takes time, or produces useful findings. Skip reviews, code audits, routine research and debugging, status reports, test runs, minor UI polish, formatting or dependency bumps, and docs-only changes. The maintainer may explicitly request a record for an otherwise non-qualifying task. When in doubt, do not write one.

Workflow

1. New entry — plan first, before coding
  1. Check the index for a related entry; an in-frame follow-up belongs there (section 2).
  2. Pick the next number: ls docs-ai/ | sort, take the highest NNN + 1 (three digits).
  3. Create docs-ai/NNN-<kebab-slug>/000-plan.md from the template below with status Planned. Write it as part of planning, not as an afterthought.
  4. Implement the work within the user's authorization. A plan or this skill never authorizes a commit, push, or other remote mutation.
  5. Write 001-action.md: what actually happened, chronologically, with commit refs, the resulting key files, validation you actually performed, and deviations from the plan. Flip the plan status to Implemented only when the intended scope is complete; use In progress for partial delivery. When committing is authorized, commit the docs with the change.
  6. Add or refresh the entry's row in docs-ai/README.md.
2. Follow-up on an existing entry (in-frame fix or extension)
  1. Add the next-numbered file in the folder, for example 002-<topic>.md (template below).
  2. At the end of 000-plan.md's Amendments section append: - Updated YYYY-MM-DD: <one line> — see [002-<topic>.md](002-<topic>.md).
  3. If the follow-up invalidates part of the plan or action text, correct that text in place (keep it truthful) and note the correction in the amendment.
  4. Work delivered in several slices: each slice adds its own 00N-<slice>.md amendment, starting at 002. Write 001-action.md once, when the last slice lands (or the entry is superseded), summarizing the slices.
Show full SKILL.md (272 more words)Show less
3. Large pivot / redesign

If the change replaces the entry's approach rather than patching it, open a NEW numbered entry, cross-link both directions, and mark the old plan Superseded by [NNN-new-slug](../NNN-new-slug/000-plan.md).

Templates

000-plan.md
markdown
# NNN — <Title>: Plan

| Field | Value |
| --- | --- |
| Status | Planned \| In progress \| Implemented \| Superseded by <link> |
| Anchor date | YYYY-MM-DD |
| Origin | <who asked, or what triggered the work> |
| Related | [NNN-other](../NNN-other/000-plan.md), `docs/...` |

## Background
The product problem and its context; for a fix, the observed symptom.

## Goals
Bullets. Add a **Non-goals** subsection when scope exclusion is a real decision.

## Design / approach
The intended approach; name the key types and files it touches.

## Alternatives & decisions
| Choice | Reason and alternative |
| --- | --- |
Options considered and why the chosen one won. Record decisions, not just designs.

## Validation and acceptance criteria
What must be checked, and how (focused tests, simulator run, manual checks).

## Amendments
None yet.
001-action.md
markdown
# NNN — <Title>: Action record

## Timeline
| Date | Change | Ref |
| --- | --- | --- |

## Outcome and current state (as of YYYY-MM-DD)
What exists in code now, followed by key source files as repo-relative paths.

## Deviations from plan
Where reality diverged from 000-plan.md, or "None known."

## Validation
Commands and checks actually run, with results. Omit nothing that failed.

## Open questions
Unverified claims, oddities worth revisiting, or "None."
Amendment (002+)
markdown
# NNN.00M — <Topic>

## Context
Why this follow-up happened.

## Change
What was decided and done, and any rejected alternative.

## Refs
Commits, files, tests. Add "## Current state" when useful.

Writing rules

  • English, factual, RFC-ish; prefer tables for timelines and decisions. Keep entries long enough to be useful and short enough to be read.
  • Every repo-relative path you write must exist (verify before writing); label any proposed new path as proposed. Facts you can't verify belong under Open questions, not in prose.
  • Never invent a rationale from code alone. If the reason for a choice is unknown, say so; if you infer it from the implementation, label it as inference.
  • Do not state build or test results you didn't produce. Old passing tests and checked task lists are not fresh validation.
  • Reference commits by short hash in backticks; dates come from Git. Cross-link sibling entries with relative links.
  • Numbered files are history. Non-numbered files inside an entry folder (for example 006-.../broadcast-availability-policy.md, 005-.../requirements/) are living documents: update them in place when the contract they describe changes, and link them instead of duplicating their content. Files under historical/ are preserved sources; do not edit them.
  • A plan reconstructed after the work shipped must say so in its Origin field; never imply it preceded implementation. Such plans add an Evidence and recovered rationale section separating source documents, recalled discussion, and inference.
  • docs/ holds public tutorials and images; history, decisions, contracts, and runbooks belong in docs-ai/. Asset-package READMEs stay with their assets.
  • AniShelf sources live in MyAnimeList/Sources/ and DataProvider/Sources/. Use AGENTS.md for Makefile commands, test filters, simulator rules, and localization.

© samuelhe52, Apache-2.0. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

Just SKILL.md in .agents/skills/write-ai-doc of samuelhe52/AniShelf.

Open the folder on GitHubat commit 5957f72

Compare with similar skills

Write AI Doc next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Write AI Doc compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Write AI Doc this skillsamuelhe52/AniShelf140—~1.8kAutomated safety check: PassApache-2.0
Update Swiftui APIsAvdLee/SwiftUI-Agent-Skill3.7k—~1.2kAutomated safety check: PassMIT
SwiftUI Design SkillWholiver/swiftui-design-skill210—~2.8kAutomated safety check: PassMIT
Audit Xcode Security Settingssuperagents-lab/xcode27-skills338—~4.6kAutomated safety check: PassNone
iOS Marketing CaptureParthJadhav/ios-marketing-capture262—~6.1kAutomated safety check: PassMIT
Preview BuildIron-Ham/XcodePreviews152—~782Automated safety check: PassMIT

Similar skills

  • Update Swiftui APIs

    AvdLee/SwiftUI-Agent-Skill

    Scan Apple's SwiftUI documentation for deprecated APIs and update the SwiftUI Expert Skill with modern replacements.

    3.7k GitHub stars~1.2k tokensUpdated 2 days ago
    MobileAuto-check passed
  • SwiftUI Design Skill

    Wholiver/swiftui-design-skill

    Guides the agent to design distinctive SwiftUI interfaces for iOS and macOS, with six anti-generic rules, a design direction workflow and a five-dimension review.

    210 GitHub stars~2.8k tokensUpdated 5 mo ago
    MobileAuto-check passed
  • Audit Xcode Security Settings

    superagents-lab/xcode27-skills

    Audit and enable security-oriented Xcode build settings. An agent skill from superagents-lab/xcode27-skills.

    338 GitHub stars~4.6k tokensUpdated 4 mo ago
    MobileAuto-check passed
  • iOS Marketing Capture

    ParthJadhav/ios-marketing-capture

    A skill your agent uses when the user wants to automate capture of marketing screenshots for a SwiftUI iOS app across multiple locales, devices, or appearances.

    262 GitHub stars~6.1k tokensUpdated 1 mo ago
    MobileAuto-check passed
  • Preview Build

    Iron-Ham/XcodePreviews

    Build and capture SwiftUI previews for visual analysis. An agent skill from Iron-Ham/XcodePreviews.

    152 GitHub stars~782 tokensUpdated 3 mo ago
    MobileAuto-check passed
  • Flowdeck

    SwiftedMind/Tessera

    FlowDeck is REQUIRED for all Apple platform build/run/test/launch/debug/simulator/device/log/automation tasks.

    116 GitHub stars~11k tokensUpdated 5 mo ago
    MobileAuto-check passed

More from samuelhe52/AniShelf

  • Anishelf Release

    samuelhe52/AniShelf

    Prepare and publish AniShelf source-control releases through validation, release commits, annotated tags, and the main-to-release pull request.

    140 GitHub stars~2k tokensUpdated yesterday
    Auto-check passed

Works with

Categories

Questions about Write AI Doc

What does Write AI Doc do?

Create and maintain curated AniShelf docs-ai/ records for substantial features and non-trivial, decision-shaping fixes (numbered entries with 000-plan.md before implementation and 001-action.md…. Write AI Doc is an agent skill from samuelhe52/AniShelf.md after).

When should I use Write AI Doc?

Write AI Doc fits situations like: routine investigations; status reports unless the maintainer explicitly asks for a docs-ai/ record.

How do I install Write AI Doc in Claude Code?

Run `npx skills add samuelhe52/AniShelf --skill write-ai-doc -a claude-code`. Or copy the skill folder (.agents/skills/write-ai-doc in samuelhe52/AniShelf) into .claude/skills/write-ai-doc in your project. Claude Code loads it when a task matches its description.

How do I install Write AI Doc in Codex?

Run `npx skills add samuelhe52/AniShelf --skill write-ai-doc -a codex`. Or copy the skill folder (.agents/skills/write-ai-doc in samuelhe52/AniShelf) into .agents/skills/write-ai-doc in your project. Codex loads it when a task matches its description.

Can I use Write AI Doc in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add samuelhe52/AniShelf --skill write-ai-doc -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/write-ai-doc, .gemini/skills/write-ai-doc, .github/skills/write-ai-doc and .opencode/skills/write-ai-doc in your project.

What does Write AI Doc need to run?

SKILL.md names no scripts, command-line tools or credentials: Write AI Doc is instructions for the agent only.

Does Write AI Doc access the network?

SKILL.md names 1 domain. As links in the text: github.com. This is read from the text; nothing was executed.

Is Write AI Doc safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does Write AI Doc use?

Write AI Doc is published under the Apache-2.0 licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Write AI Doc use?

About 1.8k tokens (SKILL.md is roughly 7k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to Write AI Doc?

Skills that share tags, products or a category with Write AI Doc: Update Swiftui APIs (AvdLee/SwiftUI-Agent-Skill, 3.7k stars), SwiftUI Design Skill (Wholiver/swiftui-design-skill, 210 stars), Audit Xcode Security Settings (superagents-lab/xcode27-skills, 338 stars) and iOS Marketing Capture (ParthJadhav/ios-marketing-capture, 262 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Write AI Doc?

samuelhe52 (a GitHub user) maintains it in samuelhe52/AniShelf, which has 140 GitHub stars. The repository holds 2 skills in this directory. The repository was last updated on October 7, 2026.

Source: samuelhe52/AniShelf on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.