create-readme
Pass
Audited by Gen Agent Trust Hub on Oct 10, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute
gh api user --jq .loginto retrieve the current GitHub username. This is used as a logic branch to determine if a Chinese version of the README should be generated for personal repositories. - [INDIRECT_PROMPT_INJECTION]: The skill reads the entire project workspace and local reference files in the
references/directory to gather context for the README. This creates a surface for indirect prompt injection if the workspace contains files with malicious instructions, which is a common risk for documentation-generating agents.
Audit Metadata