review-pr
Pass
Audited by Gen Agent Trust Hub on Nov 13, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted pull request data including diffs, branch content, and PR metadata which creates a surface for indirect prompt injection attacks.
- Ingestion points: The skill ingests data from the output of
git diff,gh pr view, and local source files analyzed bygo testandmakeas described inSKILL.md. - Boundary markers: Absent. The instructions do not define delimiters or "ignore instructions" prompts for the agent when processing external repository content.
- Capability inventory: The skill executes several shell commands (
git,gh,go test,make) and writes a review report to a local file (.cursor/review-pr-output.md) as defined inSKILL.md. - Sanitization: Absent. There is no mention of sanitizing or escaping the diff content, branch names, or metadata before processing.
- [COMMAND_EXECUTION]: The skill instructs the agent to execute various shell commands to resolve the target branch and validate changes.
- Evidence: Command sequences for resolving diff targets using
git(diff, fetch, branch) andgh(pr view) are provided inSKILL.md(Step 1). - Evidence: Automated validation commands such as
go test ./code/go/internalandmake checkare specified inSKILL.md(Step 4). - Evidence: These commands target the vendor's repository
elastic/package-spec, which is consistent with the skill's purpose.
Audit Metadata