The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill uses bunx to fetch and execute the cc-skills-usage@0.1.1 package from the npm registry at runtime.
[COMMAND_EXECUTION]: Uses the Bash tool to execute system commands for data retrieval, specifically invoking the external cc-skills-usage utility.
[INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection by processing untrusted data from multiple sources.
Ingestion points: The skill reads conversation history (specifically the triggerMessage and userMessages fields) and external SKILL.md files from the local filesystem (~/.claude/skills/).
Boundary markers: There are no explicit instructions to the agent to treat the retrieved conversation data as data rather than instructions, nor are there delimiters defined.
Capability inventory: The skill has access to the Bash tool (which it uses for bunx), the Read tool, and the Glob tool.
Sanitization: No sanitization or filtering of the retrieved user messages or skill definitions is performed before the agent analyzes them.

analyze-skill