The Agent Skills Directory

[PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection because its core functionality involves ingesting and acting upon untrusted external data.
Ingestion points: In references/capturing-updates.md, the skill defines a routine to log external messages (emails, DMs, DMs) into the threads/ directory and update the central deal-tracker.md based on this input.
Boundary markers: The instructions lack boundary markers or specific guidance to the agent to disregard instructions potentially embedded within the ingested communications.
Capability inventory: The agent is granted extensive file-system capabilities to manage the deal repository, including creating new files, appending to logs, and performing global search-and-replace operations (using grep) to maintain data consistency across the vault.
Sanitization: There is no requirement or routine provided for sanitizing or escaping the content of external communications before they are processed and stored in the primary tracking files.

mna-advisor