The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to process external data from CRM-style exports and a local context file, which represents an attack surface for indirect instructions.
Ingestion points: SKILL.md instructs the agent to read ~/.claude/company-context.md and ingest data from JSON files (revenue_data.json, pipeline_data.json, sales_data.json) using the provided Python tools.
Boundary markers: The instructions lack specific guidance for the agent to use delimiters or to disregard potentially malicious instructions embedded within the processed data files.
Capability inventory: The skill includes and executes three local Python scripts (revenue_waterfall_analyzer.py, pipeline_coverage_calculator.py, sales_efficiency_scorer.py) and performs file read operations.
Sanitization: There is no evidence of input validation, filtering, or escaping of the external content before it is processed by the agent or the analysis scripts.

cro-advisor