posting-bitwarden-review-comments
Pass
Audited by Gen Agent Trust Hub on Oct 24, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill defines a workflow for processing code review findings and outputting them to GitHub pull requests or local files. This creates a surface for indirect prompt injection if the findings are derived from untrusted code or PR descriptions that contain malicious instructions meant to influence the review process.
- Ingestion points: Findings and review data derived from code analysis or GitHub threads.
- Boundary markers: Uses Markdown structure and
<details>HTML tags for formatting, but does not specify delimiters or "ignore previous instructions" warnings for the data being processed. - Capability inventory: Posting comments to GitHub via
mcp__github_inline_comment__create_inline_comment, adding replies viamcp__github_replies__add_reply_to_pull_request_comment, and writing toreview-inline-comments.mdin the working directory. - Sanitization: The skill lacks explicit instructions for sanitizing or escaping finding content before it is emitted to the output destinations.
Audit Metadata