The Agent Skills Directory

[COMMAND_EXECUTION] (HIGH): The skill instructs the agent to execute system-level commands with superuser privileges.
Evidence: Found sudo apt-get install postgresql postgresql-contrib and sudo systemctl start postgresql in the PostgreSQL setup section of SKILL.md.
Risk: Granting an agent sudo access allows for full system compromise, unauthorized software installation, and persistence.
[PROMPT_INJECTION] (HIGH): The skill exhibits a significant Indirect Prompt Injection attack surface (Category 8).
Ingestion points: The skill is designed to read data from external sources (MongoDB and PostgreSQL databases) via commands like db.users.find() and SELECT * FROM users.
Boundary markers: Absent. There are no instructions or delimiters provided to the agent to treat database content as untrusted or to ignore embedded instructions.
Capability inventory: The skill possesses high-risk capabilities including shell access via psql and mongosh, as well as Python script execution for migrations and backups.
Sanitization: Absent. There is no evidence of output sanitization or validation for data retrieved from the databases before it is processed or used in further commands.
Risk: An attacker could place malicious instructions inside a database field. When the agent reads this record, it could be coerced into executing arbitrary SQL, exfiltrating data, or modifying the system via the provided Python utility scripts.

databases