fix: update vendored skills to latest versions#236
Conversation
pleaseai-release-bot
Bot
added
the
dependencies
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
|
PR author is in the excluded authors list. |
🔍 Tessl Skill Review
|
| Dimension | Score | Detail |
|---|---|---|
| conciseness | ██░ 2/3 | The skill is relatively lean but includes a 'Why agent-browser' marketing section that adds no actionable value for Claude. The observability dashboard section is useful but could be tighter. The core approach of pointing to CLI-served docs is efficient, though the explanation of why it's a stub is slightly verbose. |
| actionability | ██░ 2/3 | The skill provides concrete commands for loading specialized skills and installing the tool, but the actual workflow guidance is entirely deferred to agent-browser skills get core. Claude cannot act on this skill alone — it must first run a command to get the real instructions. The install command and skills commands are copy-paste ready, but the skill itself is not self-contained enough to be fully actionable. |
| workflow clarity | ██░ 2/3 | There is a clear sequence implied: install, then run skills get core to load the actual workflow. However, the skill lacks any inline workflow for common tasks, validation steps, or error recovery guidance. The 'Start here' section provides a two-step discovery process but no actual task workflow. |
| progressive disclosure | ██░ 2/3 | The skill uses a discovery-stub pattern pointing to CLI-served content, which is a reasonable progressive disclosure approach. However, since no bundle files are provided and the actual content lives entirely outside the skill file (served by the CLI), there's no way to verify the referenced content exists or is well-structured. The specialized skills section is well-organized with clear one-level-deep references, but the core content is essentially empty — it's all deferred. |
Overall: This skill is essentially a pointer/stub that defers all substantive content to CLI-served documentation via agent-browser skills get core. While the approach of serving version-matched docs from the CLI is clever, it means the SKILL.md itself provides almost no actionable guidance — Claude must execute a command before it can do anything useful. The 'Why agent-browser' section is marketing fluff that wastes tokens, and the observability dashboard section is the only piece of genuinely useful inline information.
Suggestions:
- Include at minimum a condensed quick-start workflow inline (e.g., start session, navigate, snapshot, interact, close) so Claude can begin working without first running
skills get core - Remove the 'Why agent-browser' section entirely — it's marketing content that provides no actionable guidance and wastes context window tokens
- Add a brief troubleshooting section for common issues (e.g., Chrome not found, CDP connection failures) so Claude doesn't need to load external content for basic error recovery
- Include at least one concrete end-to-end example (e.g., navigate to URL, take screenshot) so the skill is minimally self-contained
plugins/agent-browser/agent/skills/agent-browser/SKILL.md
⚠️ Error: - Reviewing skill...
✘ Skill validation failed
plugins/ai-sdk/.agents/skills/ai-sdk/SKILL.md
Review Details
Review Details
| Dimension | Score | Detail |
|---|---|---|
| conciseness | ██░ 2/3 | The skill is mostly efficient but has some redundancy — the model fetching instruction appears in both the main body (point 7) and the AI Gateway reference, and the 'Critical: Do Not Trust Internal Knowledge' framing is somewhat verbose. However, most content earns its place given the complexity of API migration guidance. |
| actionability | ███ 3/3 | The skill provides concrete, executable commands throughout: specific grep commands for searching docs, curl commands for fetching models, clear file paths to check, and explicit instructions for package installation. The references contain fully executable TypeScript code examples with correct/incorrect patterns. |
| workflow clarity | ███ 3/3 | The skill presents a clear sequenced workflow: check version → install package → search local docs → fall back to remote docs → check common errors before debugging → run typecheck after changes. The 'When Typecheck Fails' section provides an explicit feedback loop (check common-errors → search source → search remote). The prerequisites establish a clear validation checkpoint. |
| progressive disclosure | ███ 3/3 | The main SKILL.md serves as a concise overview with well-signaled, one-level-deep references to four focused reference files (common-errors.md, ai-gateway.md, type-safe-agents.md, devtools.md). Each reference is appropriately scoped and clearly linked from relevant sections in the body. The bundle structure matches the references perfectly. |
Overall: This is a well-structured skill that effectively guides Claude through AI SDK v7 development with clear workflows, actionable commands, and excellent progressive disclosure via focused reference files. The main weakness is minor redundancy between the body and reference files (e.g., model fetching instructions) and slightly verbose framing around the 'do not trust internal knowledge' section, though this emphasis is arguably justified given the migration-heavy context.
plugins/ai-sdk/agent/skills/ai-sdk/SKILL.md
⚠️ Error: - Reviewing skill...
✘ Skill validation failed
plugins/antfu/.agents/skills/antfu/SKILL.md
Review Details
Review Details
| Dimension | Score | Detail |
|---|---|---|
| conciseness | ███ 3/3 | The content is lean and efficient throughout. It avoids explaining basic concepts Claude already knows (e.g., what ESLint is, how TypeScript works), uses tables for compact reference, and every section delivers actionable configuration or convention without padding. |
| actionability | ███ 3/3 | The skill provides fully executable code snippets (tsconfig.json, eslint.config.mjs, package.json configs), concrete CLI commands via the ni command table, and specific tool recommendations with exact usage patterns. Everything is copy-paste ready. |
| workflow clarity | ██░ 2/3 | The skill presents conventions and tooling choices clearly but lacks explicit multi-step workflows with validation checkpoints. For example, setting up a new project doesn't have a sequenced checklist with verification steps. The content is more of a reference than a guided workflow, though for a conventions/preferences skill this is partially acceptable. |
| progressive disclosure | ███ 3/3 | Excellent progressive disclosure with a clear overview in the main SKILL.md and well-signaled one-level-deep references to five supporting files. The references table at the bottom provides clear navigation with descriptions, and all referenced paths match the actual bundle file listing. |
Overall: This is a well-structured, highly actionable skill that efficiently communicates Anthony Fu's tooling conventions. Its greatest strengths are conciseness (no wasted tokens), actionability (executable configs and commands throughout), and excellent progressive disclosure with a clean reference architecture. The main weakness is the lack of explicit multi-step workflows with validation checkpoints for processes like project setup or library publishing, though the conventions-focused nature of the skill partially mitigates this.
Suggestions:
- Consider adding a brief sequenced workflow (e.g., 'New Project Setup Checklist') with numbered steps and validation checkpoints like 'verify lint passes' or 'confirm build output exists' to improve workflow clarity.
plugins/antfu/agent/skills/antfu/SKILL.md
⚠️ Error: - Reviewing skill...
✘ Skill validation failed
plugins/chat-sdk/.agents/skills/chat-sdk/SKILL.md
Review Details
Review Details
| Dimension | Score | Detail |
|---|---|---|
| conciseness | ██░ 2/3 | The content is moderately efficient but includes some unnecessary verbosity, particularly in the long resource listings with full descriptions. The guide descriptions could be more concise since Claude can read the actual files. The adapter catalog section is reasonably lean. |
| actionability | ██░ 2/3 | The skill provides concrete file paths and import names, which is useful, but lacks executable code examples. There are no code snippets showing how to initialize an adapter, handle a message, or set up a webhook route. The guidance is mostly 'go read these docs' rather than providing copy-paste-ready patterns. |
| workflow clarity | █░░ 1/3 | There is no clear multi-step workflow for building a bot. The skill doesn't sequence steps like 'scaffold → configure adapter → handle events → deploy' with validation checkpoints. It's essentially a resource directory rather than a workflow guide. |
| progressive disclosure | ███ 3/3 | The skill effectively serves as an overview/index pointing to detailed guides, templates, and documentation in node_modules paths. References are one level deep and clearly signaled with descriptive summaries. The structure is well-organized with clear sections. |
Overall: This skill functions well as a resource directory and navigation guide for Chat SDK, with excellent progressive disclosure pointing to detailed guides and templates. However, it lacks actionable code examples and clear workflow sequences that would help Claude actually build a bot without needing to read multiple external files first. The content would benefit significantly from a quick-start code example and a sequenced workflow.
Suggestions:
- Add a concrete quick-start code example showing minimal bot setup (e.g., initializing an adapter, handling a message event, and responding) so Claude can act without reading external docs first.
- Add a sequenced workflow section like '1. Scaffold with create-chat-sdk → 2. Configure adapter → 3. Set env vars → 4. Handle events → 5. Test locally → 6. Deploy' with validation checkpoints.
- Trim the guide descriptions to one-line summaries since Claude can read the full guide files when needed, saving significant token budget.
plugins/chat-sdk/agent/skills/chat-sdk/SKILL.md
⚠️ Error: - Reviewing skill...
✘ Skill validation failed
plugins/emulate/.agents/skills/emulate/SKILL.md
Review Details
Review Details
| Dimension | Score | Detail |
|---|---|---|
| conciseness | ██░ 2/3 | The skill is mostly efficient and provides useful reference tables, but the extensive per-service YAML config examples are quite verbose — many follow identical patterns (users, oauth_clients, redirect_uris) that could be shown once with a note that other services follow the same shape. The base URL priority explanation is repeated in multiple places. However, it avoids explaining basic concepts Claude already knows. |
| actionability | ███ 3/3 | The skill provides fully executable CLI commands, copy-paste-ready TypeScript code for programmatic usage and test setup, complete YAML config examples with realistic values, and concrete environment variable names. Every section gives specific, actionable guidance. |
| workflow clarity | ██░ 2/3 | The skill clearly covers individual operations (starting services, configuring seed data, test setup/teardown) but lacks explicit multi-step workflow sequences with validation checkpoints. For example, there's no 'verify the emulator is running' step after startup, no troubleshooting guidance if a service fails to start, and no validation that the app is correctly pointed at the emulator. |
| progressive disclosure | ██░ 2/3 | The skill references external files ('See the next skill' for Next.js integration, 'See each service's skill for SDK-specific override instructions') which is good progressive disclosure, but the main file itself is quite long with inline config examples for every service that could be split into per-service reference files. The architecture section at the end hints at a package structure but doesn't link to per-service documentation. No bundle files are provided to support the references. |
Overall: This is a comprehensive and highly actionable skill that provides concrete CLI commands, programmatic API usage, test setup patterns, and detailed configuration examples. Its main weaknesses are verbosity in the configuration section (repetitive per-service YAML patterns) and lack of explicit validation/verification steps in workflows. The content would benefit from consolidating repetitive config examples and adding startup verification steps.
Suggestions:
- Consolidate the per-service YAML config examples: show one service in full detail, then note that others follow the same pattern with a brief table of service-specific field names.
- Add a verification step after startup (e.g., 'Confirm services are running: curl http://localhost:4001/health') and troubleshooting guidance for common failures.
- Remove the duplicated base URL priority explanation — it appears in both the CLI Options section and the Programmatic API Options section. Define it once and reference it.
plugins/emulate/.agents/skills/slack/SKILL.md
Review Details
Review Details
| Dimension | Score | Detail |
|---|---|---|
| conciseness | █░░ 1/3 | Extremely verbose at ~500+ lines. Massive repetition of curl examples that follow identical patterns — Claude can extrapolate from 2-3 examples. The full scope list is repeated multiple times (seed config, auth section). The seed config YAML alone is ~60 lines with every scope listed. Much of this is reference documentation that Claude doesn't need inline. |
| actionability | ███ 3/3 | Every endpoint has fully executable, copy-paste-ready curl commands with proper headers, JSON bodies, and realistic placeholder values. The seed config is complete YAML. SDK integration examples are concrete and executable. |
| workflow clarity | ██░ 2/3 | The OAuth flow is clearly sequenced as a numbered list. The file upload flow (getUploadURL → upload → completeUpload) is shown step-by-step. However, there are no validation checkpoints or error recovery steps — e.g., no guidance on what to do if auth fails, if a channel doesn't exist, or how to verify the emulator is running before making calls. |
| progressive disclosure | █░░ 1/3 | This is a monolithic wall of text with no references to external files. The entire API reference (~400 lines of curl examples) is inline when it could be split into separate reference files. No bundle files exist to offload content to. The document would benefit enormously from a concise overview pointing to separate endpoint reference docs. |
Overall: The skill is highly actionable with excellent, executable curl examples for every endpoint, but it is severely over-long and monolithic. The content reads like a complete API reference manual rather than a concise skill document — scope lists are repeated, curl patterns are duplicated dozens of times, and there's no progressive disclosure structure. Token efficiency is poor; Claude could infer most endpoint patterns from 3-4 examples rather than 40+.
Suggestions:
- Drastically reduce the number of inline curl examples — show 3-4 representative patterns (post message, list with pagination, file upload flow, OAuth exchange) and note that other endpoints follow the same pattern with a brief endpoint table.
- Extract the full API endpoint reference into a separate REFERENCE.md file and link to it from the main skill, keeping only a quick-start section and common patterns inline.
- Remove duplicate scope lists — list supported scopes once (ideally in a reference file) rather than repeating them in the auth section, seed config, and oauth_apps config.
- Add validation/verification steps: e.g., 'Verify the emulator is running with
curl $BASE/api/auth.test' before proceeding, and mention common error responses and how to handle them.
plugins/emulate/agent/skills/emulate/SKILL.md
⚠️ Error: - Reviewing skill...
✘ Skill validation failed
plugins/emulate/agent/skills/slack/SKILL.md
⚠️ Error: - Reviewing skill...
✘ Skill validation failed
To improve your score, point your agent at the Tessl optimization guide. Need help? Jump on our Discord.
Feedback
Report issues with this review at tesslio/skill-review, or send private feedback from your terminal with tessl feedback.
Up to standards ✅🟢 Issues
|
|
1 participant
Automated refresh of vendored skills.sh skills to their latest upstream versions.
Updated lock directories:
plugins/agent-browserplugins/ai-sdkplugins/antfuplugins/chat-sdkplugins/emulateplugins/eveplugins/nuxtplugins/pnpmplugins/portlessplugins/react-nativeplugins/reactplugins/skill-optimizerplugins/slack-agentplugins/turborepoplugins/vitestplugins/web-designplugins/workflow-sdkGenerated by
.github/workflows/update-skills.yml→scripts/update-skills.ts.