The Agent Skills Directory

[COMMAND_EXECUTION]: The skill's operational logic incorporates the execution of local scripts for codebase analysis. For instance, the rule defined in rules/complexity-breakdown.md references the use of ./scripts/analyze-codebase.sh to calculate metrics for task decomposition.
[COMMAND_EXECUTION]: In SKILL.md, a PreToolUse hook is configured to run a local script ${CLAUDE_PLUGIN_ROOT}/hooks/bin/run-hook.mjs whenever a Read tool call is made, which is a standard pattern for managing assessment baselines and state within this framework.
[PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection as it ingests untrusted codebase data and interpolates it into prompts for specialized sub-agents. While it uses boundary markers like 'Scope Constraint' headers, it is inherently susceptible to instructions embedded in analyzed files. This is noted as a low-risk architectural attribute necessary for its primary function.

assess