From project-tools
Analyze each skill in the marketplace and recommend whether it is better suited as a skill or command
How this skill is triggered — by the user, by Claude, or both
Slash command
/project-tools:audit-skill-typeThis skill is limited to the following tools:
The summary Claude sees in its skill listing — used to decide when to auto-load this skill
Scan every `SKILL.md` in the `plugins/` directory and evaluate whether each is better implemented as a **skill**
Scan every SKILL.md in the plugins/ directory and evaluate whether each is better implemented as a skill
(rich markdown instructions with multi-step workflows) or a command (lightweight executable action). Output a
recommendation table.
In the Claude Code plugin system, skills and commands serve different purposes:
Most items in this marketplace are skills. This audit identifies any that would be better served as commands.
Evaluate each skill against five complexity signals:
| Signal | Favors Skill | Favors Command |
|---|---|---|
| Step count | 3+ steps with branching or conditional logic | 1–2 simple linear steps |
| Context needs | Requires domain knowledge or codebase understanding | Needs only file paths or simple args |
| Tool variety | Uses 3+ different tools from allowed-tools | Uses mainly Bash or Write alone |
| Output complexity | Produces structured analysis, reports, multi-file changes | Produces a single file or simple output |
| Bundled resources | References or would benefit from scripts/, references/, assets/ | None needed |
Discover all skills: Glob for plugins/*/skills/*/SKILL.md to find every skill in the marketplace.
Read each skill: For each SKILL.md, read the full file. Parse:
name and allowed-tools fields from the YAML frontmatter## StepsEvaluate each signal for the skill:
## Steps. 3+ with branching favors skill.allowed-tools. 3+ favors skill; 1–2 or absent favors command.Determine recommendation: Apply the scoring rules above to produce a recommendation (Skill or Command) and confidence level (High, Medium, Low) for each.
Build the recommendation table: Output a markdown table with these columns:
| # | Skill | Plugin | Recommendation | Confidence | Reasoning |
Summarize: After the table, output:
npx claudepluginhub mattbobambrose/mattbobambrose-claude-skills --plugin project-toolsInternal development tool that performs a comprehensive quality audit of all marketing skills and agents. Reviews description overlap, format consistency, feature adoption, and structural patterns. Use when the user says "audit skills", "skill quality review", "check skill health", "are my skills consistent", "monthly skill review", or before releasing a new version of the skills repo.
Migrates legacy .claude/commands/ files to .claude/skills/ directories in Claude Code repos. Actions: discover (find), analyze (map), migrate, audit, validate, fix.
Audits Claude Code skills for structure compliance, triggering accuracy, instruction quality, and best practices. Scores 0-100 with prioritized improvement recommendations.