jaan-issue-review
Pass
Audited by Gen Agent Trust Hub on Feb 24, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes
ghandgitcommands to manage GitHub issues, branches, and PRs. These operations are directed at the developer's specific repository (parhumm/jaan-to). - [PROMPT_INJECTION]: The skill fetches content from external GitHub issues (
gh issue view), which introduces a surface for indirect prompt injection. A malicious issue could attempt to influence the automated planning phase. - Ingestion points: External data enters the workflow via GitHub issue bodies and comments in
SKILL.md. - Boundary markers: Content is summarized and parsed, but no explicit technical delimiters are used to isolate untrusted data within the prompt context.
- Capability inventory: The agent has permissions to edit/write files and execute shell commands, which could be exploited if an injection is successful.
- Sanitization: The risk is effectively managed through a dedicated 'Security Review' step (Step 1.7) that evaluates plans for injection and privilege escalation, a security validation script (Step 5.5), and mandatory user approval 'Hard Stops' before any implementation or merging takes place.
Audit Metadata