feat: improve skill scores across all five skills

[rohan-tessl]

Hey @cpoepke 👋

I ran your skills through tessl skill review at work and found some targeted improvements.

Here's the full before/after:

Skill	Before	After	Change
config	55%	99%	+44%
help	66%	95%	+29%
chat	70%	94%	+24%
start	68%	90%	+22%
install	68%	81%	+13%

This PR is intentionally scoped to 5 skills to keep it reviewable. The remaining 8 skills (louder, mute, personality, quieter, stop, teammate, unmute, volume) already score well (71–86%) and can be improved in follow-ups or via automated review on future PRs.

Changes summary

All 5 skills:

• Added explicit "Use when..." clauses with natural trigger terms to improve skill discoverability
• Expanded frontmatter descriptions with concrete capabilities and trigger scenarios
• Ensured descriptions use quoted string format

config (55% → 99%):

• Added specific settings listed in description (voice, mic gain, audio device, TTS)
• Added verification step after config changes (claude-talk config KEY)
• Replaced say -v '?' with claude-talk tts voices for consistency
• Tightened both view and edit workflows

help (66% → 95%):

• Added missing commands to the table (louder, quieter, volume)
• Consolidated personality subcommands into a single reference line (removing redundancy)
• Updated "Barge-In" references to "Interrupt" to match codebase rename
• Added link to interrupt setup guide for progressive disclosure
• Added lsof command to troubleshooting

chat (70% → 94%):

• Added text-to-speech, speech, and transcription trigger terms
• Trimmed unnecessary introductory text and prerequisites section
• Made error message more actionable with specific next steps

start (68% → 90%):

• Added voice mode, hands-free coding, speak instead of type trigger terms
• Consolidated migration check into a single concise paragraph
• Added server status verification step after startup
• Reduced repeated virtualenv activation instructions
• Tightened personality adoption and teammate messaging sections

install (68% → 81%):

• Added specific technology names (Python venvs, whisper.cpp, Kokoro TTS) to description
• Added setup, configure, and reconfigure trigger terms

Honest disclosure — I work at @tesslio where we build tooling around skills like these. Not a pitch - just saw room for improvement and wanted to contribute.

Want to self-improve your skills? Just point your agent (Claude Code, Codex, etc.) at this Tessl guide and ask it to optimize your skill. Ping me - @rohan-tessl - if you hit any snags.

Thanks in advance 🙏

[cpoepke]

@rohan-tessl your agent messed up the PR. It created two PR's. One at the right repo and one in a foreign repo. 😅

[almeidazs]

@rohan-tessl thanks for the PR!!

A few fixes are needed before this is ready to merge:

• Please fix the new REST example in send-sms/SKILL.md. It currently uses https://api.rewrite.com/messages, but the rest of the repo documents https://api.rewritetoday.com/v1/messages.
• Please align the normalized output example in rewrite-inbound/SKILL.md with rewrite-inbound/references/payload-contract.md. Right now the skill introduces a different shape (webhookEventId, rewriteMessageId, nested contact) than the reference contract (eventId, messageId, contact, contactId, analysis, raw).
• Minor consistency issue: agent-sms-inbox/SKILL.md points to references/installation.md for SDK types, but that file does not actually define the payload/schema types.

The direction is good overall, especially the added quick-starts and the tighter guidance around retries/policy gates. After those fixes, this should be in much better shape to merge.

[almeidazs]

Thanks for the PR. Before this can be merged, a few changes need to be addressed:

• Fix the new REST example in send-sms/SKILL.md. It currently uses https://api.rewrite.com/messages, but the rest of the repo documents https://api.rewritetoday.com/v1/messages.
• Align the normalized output example in rewrite-inbound/SKILL.md with rewrite-inbound/references/payload-contract.md. Right now the skill introduces a different shape (webhookEventId, rewriteMessageId, nested contact) than the reference contract (eventId, messageId, contact, contactId, analysis, raw).
• Minor consistency issue: agent-sms-inbox/SKILL.md points to references/installation.md for SDK types, but that file does not actually define the payload/schema types.

The overall direction is useful, especially the added quick starts and more concrete retry/policy guidance. After those are fixed, this should be ready for another pass.