skills/pcode-xyz/simple-skills/do-db/Gen Agent Trust Hub

do-db

Warn

Audited by Gen Agent Trust Hub on Aug 27, 2026

Risk Level: MEDIUMDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTIONCOMMAND_EXECUTION
Full Analysis
  • [DATA_EXPOSURE_AND_EXFILTRATION]: The skill is designed to access the sensitive file path .env to retrieve database connection parameters including host, port, username, and password for database initialization.\n- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from local project files (docs/specs/data/table.sql, schema.json) which directly influence database schema creation and execution logic.\n
  • Ingestion points: SKILL.md reads table.sql, schema.json, and tech-stack-rule.md.\n
  • Boundary markers: Includes a user confirmation step (AskUserQuestion) and a security policy prohibiting destructive operations, but lacks explicit prompt delimiters or 'ignore' instructions for the processed file content.\n
  • Capability inventory: Database connection (write/create), SQL execution, and potential invocation of migration tools.\n
  • Sanitization: No explicit sanitization or validation of the SQL content or schema structure is mentioned before execution.\n- [DYNAMIC_EXECUTION]: The skill dynamically executes SQL statements and migration scripts derived from external documentation files found in the project directory.\n- [COMMAND_EXECUTION]: Depending on the project configuration, the skill may invoke external migration tools such as golang-migrate or ORM-specific CLI tools to perform database updates.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Aug 27, 2026, 02:49 PM
Security Audit — agent-trust-hub — do-db