skill-maturity Skill | Agent Skills

<skill> <name>skill-maturity</name> <description> Keeps capability growth intentional by assessing current skills, identifying gaps, and prioritizing additions. Prevents ad-hoc skill sprawl while ensuring the system evolves with project needs. </description> <assessment-dimensions> <dimension name="Coverage"> <question>What percentage of recurring agent tasks have a supporting skill?</question> <scoring>0-40% = Critical gap, 41-70% = Moderate gap, 71-90% = Good, 91-100% = Excellent</scoring> <method>List top 20 recurring tasks across agents; count how many have dedicated skill guidance.</method> </dimension> <dimension name="Depth"> <question>Do existing skills provide actionable steps, not just descriptions?</question> <scoring>Thin (<50 lines, no steps) = Weak, Adequate (50-100 lines, some steps) = Fair, Dense (100-150 lines, full protocol) = Strong</scoring> <method>Sample 5 skills; rate each for actionability, specificity, and anti-pattern coverage.</method> </dimension> <dimension name="Alignment"> <question>Do skills match current tech stack, workflows, and project goals?</question> <scoring>Outdated/misaligned = Weak, Partially aligned = Fair, Fully aligned = Strong</scoring> <method>Compare skill content against default-web-stack and repository-structure; flag mismatches.</method> </dimension> <dimension name="Consistency"> <question>Do skills follow a uniform structure and quality bar?</question> <scoring>Inconsistent formats = Weak, Mostly consistent = Fair, Uniform structure = Strong</scoring> <method>Compare skill file structures; identify outliers in format, depth, or naming.</method> </dimension> <dimension name="Integration"> <question>Do skills reference each other and connect to agents appropriately?</question> <scoring>Isolated skills = Weak, Some cross-refs = Fair, Well-integrated = Strong</scoring> <method>Check for integration or references sections; verify links are valid.</method> </dimension> </assessment-dimensions> <gap-identification> <sources> <source>Self-improvement logs: recurring friction pointing to missing guidance.</source> <source>Agent failures: tasks where agents struggled due to lack of skill support.</source> <source>New project needs: tech or workflow changes requiring new capabilities.</source> <source>Industry standards: common practices not yet codified (observability, accessibility, etc.).</source> </sources> <prioritization-criteria> <criterion name="Frequency">How often is this capability needed? (Daily/Weekly/Monthly/Rarely)</criterion> <criterion name="Impact">What's the cost of not having it? (Blocks work / Slows work / Minor friction)</criterion> <criterion name="Effort">How hard is it to create? (S: 1-2h, M: 2-4h, L: 4h+)</criterion> <criterion name="Dependencies">Does it require other skills or changes first?</criterion> </prioritization-criteria> <priority-formula>Priority = (Frequency × Impact) / Effort. Adjust for dependencies.</priority-formula> </gap-identification> <skill-quality-checklist> <item>Has clear description (2-3 sentences, not vague).</item> <item>Defines philosophy/principles (why, not just what).</item> <item>Provides actionable steps or protocols (how).</item> <item>Includes anti-patterns (what not to do).</item> <item>Lists success metrics (how to measure).</item> <item>Integrates with related skills and agents.</item> <item>References relevant docs where applicable.</item> <item>100-150 lines of dense, useful content.</item> <item>No placeholders, TODOs, or thin sections.</item> </skill-quality-checklist> <assessment-protocol> <step number="1">Inventory: List all skills in .github/skills/; note line counts and structure.</step> <step number="2">Score: Rate each dimension (Coverage, Depth, Alignment, Consistency, Integration) 1-5.</step> <step number="3">Gaps: Identify missing skills from sources; apply prioritization criteria.</step> <step number="4">Quality: Run checklist against each existing skill; flag those needing improvement.</step> <step number="5">Recommend: Propose top 3 new skills and top 3 skills needing enhancement.</step> <step number="6">Report: Summarize findings with scores, gaps, and recommendations.</step> </assessment-protocol> <output-format> <section name="Summary">Overall maturity score (1-5) with brief rationale.</section> <section name="Dimension Scores">Table of dimensions with scores and notes.</section> <section name="Gap Analysis">Prioritized list of missing skills with Frequency/Impact/Effort.</section> <section name="Enhancement Needs">Existing skills requiring depth or alignment fixes.</section> <section name="Recommendations">Top 3 new skills to add, top 3 existing to enhance.</section> </output-format> <cadence> <regular>Run full assessment quarterly or after major project changes.</regular> <triggered>Run partial assessment when self-improvement logs show recurring skill gaps.</triggered> <ad-hoc>Run before adding new skills to check for overlap or sequencing issues.</ad-hoc> </cadence> <anti-patterns> <anti-pattern>Adding skills reactively without assessing existing coverage first.</anti-pattern> <anti-pattern>Creating thin placeholder skills instead of dense, actionable ones.</anti-pattern> <anti-pattern>Duplicating guidance already in another skill or agent definition.</anti-pattern> <anti-pattern>Ignoring skill quality in favor of quantity.</anti-pattern> <anti-pattern>Never revisiting or enhancing existing skills as needs evolve.</anti-pattern> </anti-patterns> <integration> <with-skill name="self-improvement">Friction logs feed gap identification.</with-skill> <with-skill name="documentation">New skills may require doc updates via architect.</with-skill> <with-agent name="architect">Architect approves new skill proposals and structural changes.</with-agent> <with-agent name="orchestrator">Orchestrator routes skill gaps to appropriate owners.</with-agent> </integration> </skill>

Agent Skills: skill-maturity

Install this agent skill to your local

Skill Files