coderabbit-incident-runbook
Pass
Audited by Gen Agent Trust Hub on Jul 21, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses the GitHub CLI (
gh api) to manage repository settings, including deleting and restoring branch protection rules for the main branch. These are high-privilege operations used specifically for the stated 'P1 Emergency' purpose of unblocking Pull Requests when the review service is unavailable. - [EXTERNAL_DOWNLOADS]: Performs a network request via
curltostatus.coderabbit.ai. This is a well-known service domain used to check the operational status of the CodeRabbit platform. - [PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection by ingesting and echoing untrusted data from GitHub API responses (such as PR titles and numbers) and local configuration files (
.coderabbit.yaml). - Ingestion points: External data enters the context through
gh apicalls for PR lists and titles, and the local.coderabbit.yamlfile. - Boundary markers: None are used when echoing PR titles or processing configuration data.
- Capability inventory: The skill has the ability to modify repository protection settings and execute shell commands.
- Sanitization: The skill uses
yaml.safe_load()when parsing configuration files, which prevents unsafe object instantiation during the data ingestion phase.
Audit Metadata