The Agent Skills Directory

[DATA_EXFILTRATION]: The skill requires access to and auditing of highly sensitive database configuration files and system tables that contain authentication rules and user metadata.
Evidence: Instructs the agent to read and flag entries in pg_hba.conf, postgresql.conf, my.cnf, mongod.conf, and the mysql.user table.
PII Exposure: Explicitly directs the agent to scan for plaintext sensitive data in database columns, specifically targeting password, ssn, and credit_card patterns.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from databases and application source code, creating a potential attack surface for indirect prompt injection.
Ingestion points: Reads authentication configuration files, system roles tables, database query results, and application source code via the Grep tool (e.g., in files like scripts/database_scan.py and scripts/config_validator.py).
Boundary markers: The instructions lack explicit delimiters or "ignore embedded instructions" warnings when processing these external data sources.
Capability inventory: The skill is granted Bash, Write, and Edit permissions, providing significant execution and modification capabilities that could be targeted by injected instructions.
Sanitization: There is no mention of sanitizing, escaping, or validating the external content before it is processed by the agent.

scanning-database-security