resolve-ai-pr-reviews
Pass
Audited by Gen Agent Trust Hub on Apr 9, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses the GitHub CLI (
gh) and bash scripting to automate PR management. - It executes
gh apicalls for REST and GraphQL queries to fetch metadata and modify thread status. - It uses bash control flow (for-loops) to process and resolve multiple review threads sequentially.
- [EXTERNAL_DOWNLOADS]: The skill fetches data from the GitHub API, which is a well-known and trusted service.
- It retrieves pull request comments and issue comments for analysis.
- It queries GraphQL nodes to identify unresolved review threads.
- [PROMPT_INJECTION]: The skill contains an indirect prompt injection surface (Category 8) as it processes data from external sources.
- Ingestion points: The agent fetches PR comments authored by 'coderabbitai' and 'gemini' using the GitHub API in
SKILL.md. - Boundary markers: Absent. There are no explicit markers or instructions to the agent to treat the fetched comment bodies as untrusted data or to ignore instructions embedded within them.
- Capability inventory: The skill's workflow includes high-capability actions such as implementing code changes, running tests/linting commands, and performing write operations via the GitHub API (resolving threads and posting new comments).
- Sanitization: Absent. The agent is instructed to use the bot feedback directly to guide its code fixes without validation or sanitization of the input text.
Audit Metadata