sqlite-ops

Pass

Audited by Gen Agent Trust Hub on Apr 12, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill provides instructions and examples for using the Bash tool to interact with SQLite databases via the sqlite3 command-line interface (e.g., .tables, .schema, VACUUM). This is documented as a standard development workflow for managing local persistence.
  • [PROMPT_INJECTION]: Multiple reference files contain insecure code patterns that are vulnerable to SQL injection, potentially allowing data to influence database logic or structure.
  • Ingestion points: The patterns are designed to process data from SQLite database files and JSON payloads (e.g., in references/async-patterns.md and references/migration-patterns.md).
  • Boundary markers: The provided code does not use delimiters or instructions to prevent the agent from obeying embedded instructions in the processed data.
  • Capability inventory: The skill is granted Read, Write, and Bash tool access, which allows for file system manipulation and command execution.
  • Sanitization: The patterns lack sanitization or validation for dynamically interpolated strings. Specifically, the update_item function builds a SQL statement by interpolating dictionary keys directly into the SET clause, and the query_json_field and update_json_field functions interpolate the field parameter directly into json_extract and json_set paths.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 12, 2026, 08:39 PM
Security Audit — agent-trust-hub — sqlite-ops