The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze external, potentially untrusted documents such as Letters of Intent (LOIs), term sheets, and legal contracts.
Ingestion points: Sensitive documents and data estates are identified for analysis in SKILL.md.
Boundary markers: There are no instructions to utilize delimiters or to ignore potential instructions embedded within the target documents.
Capability inventory: The skill calls for the execution of external Python scripts (term_sheet_analyzer.py, contract_risk_scanner.py, ai_training_data_audit.py, data_asset_valuator.py) to process this data.
Sanitization: No sanitization or validation logic is defined for the content extracted from the processed files.- [COMMAND_EXECUTION]: The skill instructions explicitly direct the agent to run Python scripts found in relative sibling paths (../general-counsel-advisor/scripts/ and ../chief-data-officer-advisor/scripts/). This creates a dependency on executable code external to the current skill folder.- [METADATA_POISONING]: The updated field in the skill metadata is set to a future date (2026-03-05), which constitutes inconsistent or placeholder metadata.

ma-playbook