setup
Pass
Audited by Gen Agent Trust Hub on Nov 13, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTIONPRIVILEGE_ESCALATION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill downloads a pinned version of its CLI tool from the vendor-owned domain
https://cli.coderabbit.aiwhen the delivery mode is set to download. This process includes a security requirement to verify SHA-256 checksums for both the archive and the extracted executable before use. - [COMMAND_EXECUTION]: The skill executes its packaged CLI to inspect repositories, generate review settings, and apply configurations. It also utilizes standard system utilities like
curl,unzip, andsha256sumduring the installation phase for the runtime. - [DYNAMIC_EXECUTION]: During the configuration process, the skill generates a temporary YAML file based on repository analysis and then invokes the CLI to validate and apply this configuration to the local project.
- [INDIRECT_PROMPT_INJECTION]: The skill analyzes local project files, such as
AGENTS.mdandCLAUDE.md, to provide tailored configuration recommendations. This ingestion of untrusted data is mitigated by explicit instructions to treat the file contents as data evidence rather than executable commands, and validation is performed by the CLI against a formal schema. - [PRIVILEGE_ESCALATION]: The skill requires execution permissions for its downloaded components but includes strong guardrails. It instructs the agent to respect the host's standard permission model, forbids bypassing sandbox settings, and advises against routing around denied permissions.
Audit Metadata