agentleFS
BigDawnGhostBigDawnGhost/wenyi/.agents/skills/release-notes/SKILL.md

release-notes

Skill2.9k starsChanged yesterdayMITView on GitHub

Draft copy-ready English and Chinese Wenyi release notes from the changes since the previous stable version. Use when preparing GitHub Release descriptions, summarizing a release range, or standardizing bilingual changelog formatting.

We split this skill into the rules it gives an agent

26 rules. 0 are common, 21 are distinctive, and 5 are very specific to this project.

  • 0

    Common

    Most of the 152 edited versions we read keep these.

  • 21

    Distinctive

    This author's own choices. Worth a look.

  • 5

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Default to equivalent English and Simplified Chinese sections that Wenyi Desktop can select by interface language.

    line 12
  2. Distinctive:

    If the target is itself a stable release tag, compare against the stable release before it.

    line 32
  3. Distinctive:

    Prefer the target's first-parent release line when finding the previous release, but inspect all commits in the resulting range for changes.

    line 35
  4. Distinctive:

    When GitHub publication status is available, distinguish published stable releases from drafts/prereleases.

    line 38
  5. Distinctive:

    If tags/history are missing, the explicit base is not an ancestor, or several release lines make the previous version ambiguous, ask for …

    line 46
  6. Distinctive:

    If the verified range has no effective changes, say so rather than inventing release highlights.

    line 49
  7. Distinctive:

    Read applicable repository instructions and narrow subsequent diffs to the components whose behavior needs confirmation.

    line 66
  8. Distinctive:

    Do not claim universal compatibility, performance gains, translation-quality improvements, data preservation, security guarantees, CI …

    line 83
  9. Distinctive:

    Use real PR numbers, issue links and contributor credits only when verified.

    line 88
  10. Distinctive:

    Do not infer authorship or credit from a display name in the conversation.

    line 89
  11. Distinctive:

    Keep the two languages equivalent in capabilities, scope, warnings and upgrade instructions.

    line 90
  12. Distinctive:

    Read the current sources before drafting.

    line 102
  13. Very specific:

    Start directly with ## English.

    line 107
  14. Distinctive:

    Write a short English overview, then applicable categories.

    line 109
  15. Very specific:

    Use exactly ## 中文, followed by an equivalent Chinese overview and categories.

    line 111
  16. Very specific:

    Use ### for category headings and #### for optional deeper subheadings.

    line 112
  17. Distinctive:

    Do not put other level-one/two headings or horizontal dividers inside a language section: they end that section in Desktop.

    line 113
  18. Distinctive:

    Use these categories in this order, omitting every empty category:

    line 121
  19. Very specific:

    Use concise bullets, normally one or two sentences each.

    line 131
  20. Distinctive:

    Use backticks for commands, paths, configuration keys and code identifiers.

    line 134
  21. Distinctive:

    Prefer portable Markdown without raw HTML, images or task checkboxes.

    line 135
  22. Distinctive:

    Place language-specific links and warnings inside their language section.

    line 136
  23. Distinctive:

    Before the copyable body, briefly state the verified base and target, whether this is a released version or an unreleased draft, and any …

    line 141
  24. Distinctive:

    Do not insert stale scope disclaimers such as "a new release tag has not yet been published" into the body of an existing release.

    line 149
  25. Distinctive:

    If only one output language was explicitly requested, retain its exact language heading and omit the other section and inter-language …

    line 152
  26. Distinctive:

    Check the final Markdown: language headings are exact and unique, categories match across languages, no internal divider truncates a …

    line 154

“Common” is measured: kept by at least 6 in 10 of the 152 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

Draft Wenyi release notes

Generate a factual, user-facing GitHub Release description, not a commit dump. Default to equivalent English and Simplified Chinese sections that Wenyi Desktop can select by interface language. Deliver the Markdown in the conversation for copying; do not create a changelog file unless requested.

This skill only drafts text. It does not authorize commits, tags, merges, pushes, release creation or editing, workflow dispatch, package publication, or changes to application code. Preserve existing changes and local data.

1. Establish the comparison range

  • Honor an explicit base version, target tag/commit/branch, component scope, and requested language. Otherwise use committed HEAD as the target, covering all affected Wenyi components. Do not silently substitute main, dev, a remote branch, or the highest version visible anywhere in the repository.
  • Inspect Git status, target identity, release tags and their ancestry. Resolve the base and target to commit SHAs so the comparison is reproducible.
  • Select the nearest preceding stable version on the target's release history. Stable tags are numeric releases with one to three components, optionally prefixed with v, such as v2, v1.3, or v1.2.3; exclude prerelease, development and local versions. Check ancestry, not just version/date sorting.
  • If the target is itself a stable release tag, compare against the stable release before it. Exclude every tag alias pointing at the target commit; comparing the release to itself must not produce an empty changelog.
  • Prefer the target's first-parent release line when finding the previous release, but inspect all commits in the resulting range for changes. Do not omit feature-branch commits by using a first-parent-only change inventory.
  • When GitHub publication status is available, distinguish published stable releases from drafts/prereleases. Do not use GitHub's global "latest" release if it is outside the target's ancestry. Without publication evidence, state outside the copyable body that the baseline comes from local Git tags.
  • Exclude staged, unstaged and untracked changes by default. Mention their exclusion outside the release body. Include relevant working-tree changes only when explicitly requested; call that output an unpublished working-tree draft and keep unrelated changes/private data out.
  • If tags/history are missing, the explicit base is not an ancestor, or several release lines make the previous version ambiguous, ask for the missing base or target. Do not invent a baseline, release number, date, or publication state. If the verified range has no effective changes, say so rather than inventing release highlights.

Useful read-only inspections after resolving $base and $target:

git --no-optional-locks status --short
git remote -v
git rev-parse --verify "${base}^{commit}"
git rev-parse --verify "${target}^{commit}"
git merge-base --is-ancestor "$base" "$target"
git log --reverse --no-merges --format='%h %s' "$base..$target"
git diff --stat "$base" "$target"
git diff --name-status "$base" "$target"

These commands are starting points, not sufficient evidence for every claim. Read applicable repository instructions and narrow subsequent diffs to the components whose behavior needs confirmation.

2. Turn evidence into release highlights

  • Cross-check commit/PR descriptions against the net diff, relevant code, tests and current documentation. A reverted change is not a release feature. Deduplicate repeated fixes, follow-up commits and merge descriptions.
  • Describe the user-visible result and affected platform or workflow. Keep Desktop-only capabilities separate from Web, CLI and shared Core behavior. State constraints such as supported installers or manual setup when relevant.
  • Group related changes into meaningful highlights instead of listing every file, commit, dependency bump, test addition or internal refactor. Include development/documentation changes only when useful to the intended audience.
  • Identify evidence-backed upgrade actions: configuration changes, migration requirements, cache invalidation, credential handling, compatibility changes, extra model calls/cost, or changed defaults. Preserve important caveats.
  • Do not claim universal compatibility, performance gains, translation-quality improvements, data preservation, security guarantees, CI success, signing or platform testing without supporting evidence. Existing tests are not proof that they were run. Drafting notes does not require running the full test suite or contacting models, parser services, or private project data.
  • Use real PR numbers, issue links and contributor credits only when verified. Do not infer authorship or credit from a display name in the conversation.
  • Keep the two languages equivalent in capabilities, scope, warnings and upgrade instructions; preserve identifiers, command names and configuration keys.

3. Use the Desktop-compatible Markdown contract

Sources of truth:

  • docs/desktop.md, section "Application updates".
  • docs/zh/desktop.md, section "应用更新".
  • apps/desktop/frontend/src/localizedReleaseNotes.ts, when present, and its affected regressions in apps/desktop/frontend/tests/desktop-updates.spec.ts.

Read the current sources before drafting; do not assume a historical template still matches the application's parser.

Required structure:

  1. Start directly with ## English; do not add a version title or bilingual preface above it. GitHub's Release title is separate.
  2. Write a short English overview, then applicable categories.
  3. Insert one standalone --- between the two language sections.
  4. Use exactly ## 中文, followed by an equivalent Chinese overview and categories.
  5. Use ### for category headings and #### for optional deeper subheadings. Do not put other level-one/two headings or horizontal dividers inside a language section: they end that section in Desktop.
  6. Optionally append a shared comparison link after a final standalone ---. Derive its public repository URL from a verified source remote, never the local backlink. Pin the target to an existing release tag or resolved commit SHA, not a guessed future tag or moving main/dev ref. Omit the link if its repository or range cannot be established.

Use these categories in this order, omitting every empty category:

EnglishSimplified Chinese
New features新增功能
Improvements改进
Fixes问题修复
Development and documentation开发与文档
Upgrade notes升级说明

Use concise bullets, normally one or two sentences each. Prefer - **Short label:** Result and important scope. and its Chinese equivalent - **简短标签:** 行为变化及重要适用范围。. Use backticks for commands, paths, configuration keys and code identifiers. Prefer portable Markdown without raw HTML, images or task checkboxes. Place language-specific links and warnings inside their language section; only the optional comparison footer is shared.

4. Deliver one copyable release body

  • Before the copyable body, briefly state the verified base and target, whether this is a released version or an unreleased draft, and any excluded local changes or evidence limitations. Keep this process context in the user's conversation language, outside the Release description.
  • Return the complete body in one fenced markdown code block, using a longer outer fence if the body contains code fences. It must be ready to paste, with no placeholders, editing instructions, analysis, test transcripts, Delta worktree links, or unsupported assertions.
  • Do not insert stale scope disclaimers such as "a new release tag has not yet been published" into the body of an existing release. Do not print alternate versions of the same body unless requested.
  • If only one output language was explicitly requested, retain its exact language heading and omit the other section and inter-language divider.
  • Check the final Markdown: language headings are exact and unique, categories match across languages, no internal divider truncates a section, the overview and bullets match the verified range, and any comparison link is real.

Formatting skeleton only; replace all placeholders and omit empty categories:

## English

<Brief summary of the release's main user-facing changes.>

### New features

- **<Feature>:** <What users can now do, including platform scope.>

### Fixes

- **<Fix>:** <The corrected behavior and affected workflow.>

### Upgrade notes

- <Only upgrade actions or caveats supported by the changes.>

---

## 中文

<与英文等价的简短更新概述。>

### 新增功能

- **<功能>:** <新增能力及平台适用范围。>

### 问题修复

- **<修复>:** <修正后的行为及受影响的流程。>

### 升级说明

- <仅列出有依据的升级操作或注意事项。>

Download now (ZIP)

Trending

Quiet since July.thin dataabout the whole repository

Dec 2025: 0 mentionsJan 2026: 0 mentionsFeb 2026: 0 mentionsMar 2026: 0 mentionsApr 2026: 0 mentionsMay 2026: 0 mentionsJun 2026: 0 mentionsJul 2026: 1 mentionAug 2026: 0 mentionsSep 2026: 0 mentionsOct 2026: 0 mentionsNov 2026: 0 mentions
Dec 2025Jun 2026Nov 2026
Mentions a month of the BigDawnGhost/wenyi repository that our weekly crawl found: only 1 in all, so this is a hint, not a trend. There is no adoption data for this file yet.

Who's talking about BigDawnGhost/wenyi

The biggest 1 of 1 mention linking to the repository that our crawl found. Each line is their own headline.

  • Other versions166 projects use a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 166

    1. tw93/Mole★ 74ksame name
    2. makeplane/plane★ 62ksame name
    3. tabler/tabler★ 42ksame name
    4. pnpm/pnpm★ 37ksame name
    5. quasarframework/quasar★ 27ksame name
    6. dotnet/core★ 22ksame name
    7. zauberzeug/nicegui★ 16ksame name
    8. adobe/react-spectrum★ 16ksame name
    9. espanso/espanso★ 15ksame name
    10. formbricks/formbricks★ 13ksame name

    156 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.

  • Used alongsideProjects with it also use brainstorm-experiments-existing, brainstorm-ideas-new 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.
    • skytiger6724Skill

      brainstorm-experiments-existing

      Design experiments to test assumptions for an existing product — prototypes, A/B tests, spikes, and other low-effort validation methods. Use when validating assumptions, testing feature ideas cheaply, or planning product experiments.

      Also in 5 projects that use this

    • skytiger6724Skill

      brainstorm-ideas-new

      Brainstorm feature ideas for a new product in initial discovery from PM, Designer, and Engineer perspectives. Use when starting product discovery for a new product, exploring features for a startup idea, or doing initial ideation.

      Also in 5 projects that use this

    • 45ckSkill

      pestle-analysis

      Also in 5 projects that use this

    • launchapp-devSkill

      interview-script

      Create a structured user interview script with warm-up, core exploration, and wrap-up sections. Use when preparing for user research interviews to ensure consistent, insightful conversations.

      Also in 6 projects that use this

    • +8 more from skytiger6724
  • In this repository2 other files it gives its agents.

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.