bento-qa
Pass
Audited by Gen Agent Trust Hub on Oct 23, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes shell commands for environment inspection (lsof), repository management (git fetch, git diff), and running test suites. This is necessary for the skill's purpose but involves running code defined in the repository.
- [DYNAMIC_EXECUTION]: The agent is instructed to create and run 'scratch scripts' to test libraries and bootstrap new test harnesses if none exist in the target repository.
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted repository data that can influence its actions. (1) Ingestion points: Configuration and manifest files such as AGENTS.md, package.json, pyproject.toml, and CI configurations. (2) Boundary markers: None specified to prevent the agent from following instructions embedded within these external files. (3) Capability inventory: Includes shell execution, network operations (curl), starting background servers, and browser automation via MCP. (4) Sanitization: No explicit validation or escaping of repository-supplied configuration values is performed before they are used in command execution.
Audit Metadata