Skills

Reusable skills for coding agents. I use these across my projects and you can too. Built on the Agent Skills open specification, so they work with any agent that supports the spec.

Available Skills

You can install all skills in one go with:

npx skills add mattobee/skills

Accessibility

predicting-accessibility-risks

Identifies accessibility risks in proposed features, designs, or technical plans before implementation begins. Produces a risk assessment with affected user groups, WCAG criteria, likelihood, cost to fix later, and specific mitigations.

npx skills add mattobee/skills --skill predicting-accessibility-risks

Example prompts:

• "What accessibility risks should I watch for in this feature?"
• "Assess this design for accessibility problems before I build it"
• "What could go wrong for disabled users with this approach?"
• "Is there anything about this plan that will be expensive to fix for accessibility later?"

reviewing-accessibility

Reviews implemented UI code for WCAG Level AA compliance. Covers semantic HTML, keyboard navigation, focus management, ARIA, labels, visual contrast, and forms. For pre-implementation risk assessment, use predicting-accessibility-risks instead.

npx skills add mattobee/skills --skill reviewing-accessibility

Example prompts:

• "Review this component for accessibility"
• "Audit this page against WCAG AA"
• "Is my form handling accessible?"
• "Check whether this dialog meets accessibility requirements"

writing-accessibility-tests

Writes Playwright accessibility tests using a two-layer strategy: axe-core scans for broad automated coverage, plus targeted Playwright assertions for things axe cannot catch (accessible names, ARIA states, focus management, shadow DOM, multi-theme contrast).

npx skills add mattobee/skills --skill writing-accessibility-tests

Example prompts:

• "Add accessibility tests for this page"
• "Write axe scans for all routes in light and dark mode"
• "What accessibility test coverage am I missing?"
• "How do I test accessible names on web components with shadow DOM?"

estimating-accessibility-effort

Estimates the effort required to remediate accessibility issues. Takes audit findings, axe-core results, or backlog items and produces T-shirt-sized effort estimates with rationale, dependencies, and quick wins.

npx skills add mattobee/skills --skill estimating-accessibility-effort

Example prompts:

• "How much effort would it take to fix these accessibility issues?"
• "Size these audit findings for sprint planning"
• "Which of these accessibility fixes are quick wins?"
• "What dependencies exist between these accessibility issues?"

prioritising-accessibility-fixes

Prioritises a set of accessibility issues for remediation based on severity, user impact, and effort. Produces a scored, tiered remediation plan with batching recommendations and a concrete fix order.

npx skills add mattobee/skills --skill prioritising-accessibility-fixes

Example prompts:

• "Prioritise these accessibility issues for our next sprint"
• "What should we fix first from this audit report?"
• "Triage this accessibility backlog"
• "Which accessibility fixes give the most value for the least effort?"

fixing-accessibility-issues

Fixes accessibility issues in implemented UI code. Takes findings from audits, axe-core scans, or reviewing-accessibility and produces code changes, with pattern selection defaults for accessible names, focus management, ARIA, forms, contrast, headings, keyboard interaction, and live regions.

npx skills add mattobee/skills --skill fixing-accessibility-issues

Example prompts:

• "Fix these accessibility issues"
• "Remediate the findings from the audit"
• "How should I fix this axe-core violation?"
• "Implement an accessible modal pattern for this dialog"

General

designing-agent-teams

Designs, generates, and refines multi-agent coding teams with optimal model-to-role assignments. Produces team reasoning documents and agent instruction files that work across platforms.

npx skills add mattobee/skills --skill designing-agent-teams

Example prompts:

• "Design an agent team for this codebase"
• "Review my agent team configuration and suggest improvements"
• "Add a security specialist to my existing agent team"
• "Which model tier should I use for my tester agent?"

addressing-pr-review-comments

Works through inline review comments on a pull request: assesses each one's validity, makes the code changes that are warranted, and replies to every thread with a one-line explanation of what was done (or why it was declined).

npx skills add mattobee/skills --skill addressing-pr-review-comments

Example prompts:

• "Address the review comments on my PR"
• "Work through the feedback on this PR"
• "What do the reviewers want changed?"
• "Action the comments on PR #42"

suggesting-next-steps

Suggests prioritised next steps for a project based on git history, GitHub issues/PRs, tracking files, and project docs.

npx skills add mattobee/skills --skill suggesting-next-steps

Example prompts:

• "What should I work on next?"
• "I haven't touched this project in a week — what's the state of things?"
• "What's left to do on this feature?"
• "Help me prioritise my backlog"

License

MIT