feat: improve skill scores for codex-skills

[popey]

Hullo @am-will 👋

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
agent-browser	0%	90%	+90%
frontend-responsive-ui	0%	89%	+89%
swarm-planner	45%	89%	+44%
plan-harder	57%	89%	+32%
parallel-task	50%	81%	+31%
parallel	50%	81%	+31%
parallel-task-spark	54%	81%	+27%
super-swarm-spark	54%	81%	+27%
markdown-url	72%	93%	+21%
role-creator	76%	93%	+17%
frontend-design	59%	71%	+12%
vercel-react-best-practices	77%	86%	+9%

Six skills were already scoring well (89-100%) and didn't need changes: context7, llm-council, planner, read-github, openai-docs-skill, gemini-computer-use.

Fixed validation errors (agent-browser, frontend-responsive-ui)

• Corrected name fields to use lowercase kebab-case (agent-browser instead of Agent Browser, frontend-responsive-ui instead of Frontend Responsive Design Standards)
• Removed unsupported frontmatter keys (read_when, inline JSON metadata) from agent-browser

Expanded minimal descriptions (6 skills)

• parallel-task-spark, parallel-task, parallel, plan-harder, super-swarm-spark, swarm-planner all had descriptions that only said "triggered by explicit command" with no explanation of what they do
• Added concrete capability descriptions explaining the orchestration patterns, scheduling modes, and validation approaches each skill uses
• Added explicit "Use when..." clauses so agents can make informed selection decisions

Improved mid-scoring descriptions (4 skills)

• frontend-design: Added specific concrete actions (responsive layouts, CSS animations, forms, navigation) and more trigger terms (UI elements, websites, landing pages, dashboards)
• markdown-url: Replaced verbose multi-line description with concise single line covering what it does and when to use it
• role-creator: Added "Use when..." clause with natural language triggers (create custom agent, configure roles, set up persona)
• vercel-react-best-practices: Added specific optimization techniques and natural trigger terms (slow page loads, bundle sizes, Core Web Vitals, Lighthouse scores)

Description format standardisation

• Converted all chevron (>) YAML descriptions to quoted strings for consistent frontmatter format

Fixed step numbering

• swarm-planner had duplicate "Step 4" labels — renumbered to sequential 1-6

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 (https://docs.tessl.io/evaluate/optimize-a-skill-using-best-practices) and ask it to optimize your skill. Ping me - @popey - if you hit any snags.

Thanks in advance 🙏

[am-will]

thanks buddy, i'll check these out. appreciate it!