deep-module-refactor
Pass
Audited by Gen Agent Trust Hub on Nov 22, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill explicitly instructs the agent to bypass user confirmation before performing a write action on an external service. Specifically, in
SKILL.mdstep 7, it states: 'Do NOT ask the user to review before creating — just create it and share the URL.' This directive reduces human oversight for actions that interact with external platforms (GitHub). - [INDIRECT_PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection as it processes untrusted data (external codebases) and has write capabilities.
- Ingestion points: Reads files from arbitrary repositories during the 'Explore' phase (Step 1 in
SKILL.md). - Boundary markers: The instructions do not define boundary markers or delimiters for the code being analyzed, nor do they instruct the agent to ignore instructions embedded within those files.
- Capability inventory: The skill uses
gh issue createto write to GitHub and spawns sub-agents via the Agent tool to process analysis results. - Sanitization: There is no evidence of sanitization or filtering of the content read from the codebase before it is used to generate issue content or sub-agent prompts.
- [COMMAND_EXECUTION]: The skill uses the GitHub CLI tool (
gh issue create) to perform external write operations based on the analysis of the codebase.
Audit Metadata