coderabbit-config
Pass
Audited by Gen Agent Trust Hub on Nov 12, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [COMMAND_EXECUTION]: The skill facilitates the use of the
coderabbitCLI to perform configuration tasks. It minimizes the risk of executing unauthorized code by requiring the resolution of the CLI to a canonical absolute path and explicitly instructing the agent to reject repository-local executables, aliases, or unknown wrappers.- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external CLI operations. This potential attack surface is addressed by a comprehensive evidence chain: - Ingestion points: The agent reads data from CLI commands such as
coderabbit config --agent. - Boundary markers: It enforces a strict communication protocol (v2), verifies status flags (e.g.,
ok: true), and validates protocol versioning. - Capability inventory: The skill uses CLI execution, network access for schema validation, and temporary file writing.
- Sanitization: The instructions mandate the use of
baseHashto ensure data consistency and require the agent to present a diff for manual user approval before final saving.- [EXTERNAL_DOWNLOADS]: The skill references the official vendor documentation atdocs.coderabbit.aifor installation and update procedures. As these are established vendor resources and the instructions prohibit unauthorized updates or installations, this behavior is considered standard and safe.
Audit Metadata