The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes Python scripts located within the vendor's directory structure to perform revenue and churn analysis.
Evidence: python ${CLAUDE_PLUGIN_ROOT}/vendor/c-level-advisor/cro-advisor/scripts/revenue_forecast_model.py and python ${CLAUDE_PLUGIN_ROOT}/vendor/c-level-advisor/cro-advisor/scripts/churn_analyzer.py in SKILL.md.
[INDIRECT_PROMPT_INJECTION]: The skill ingests a <plan> argument which is then processed by automated scripts, creating a surface for potentially malicious instructions embedded in the planning data.
Ingestion points: The <plan> command argument in SKILL.md.
Boundary markers: Absent; there are no specific delimiters or instructions to the model to ignore embedded instructions within the plan data.
Capability inventory: Execution of local Python scripts using python in SKILL.md.
Sanitization: Absent; no validation or escaping of the input data is defined in the skill configuration.

cro-review