kai-surround-sound
Pass
Audited by Gen Agent Trust Hub on Sep 28, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes a Python script
scripts/quality_gates/agent_readiness_lint.pylocated within the tool's installation directory. This script is used to audit the user's domain for AI agent compatibility. - [INDIRECT_PROMPT_INJECTION]: The skill processes data from the project's codebase and external domain information to generate marketing plans, which creates a surface where external or untrusted content could influence the agent's output.
- Ingestion points: Project files (CLAUDE.md, README.md, package.json, etc.), landing pages, and external domain audits via the linter script.
- Boundary markers: Absent. The instructions do not define delimiters for untrusted data.
- Capability inventory: File system read (codebase exploration), command execution (python linter script), and file system write (workspace output).
- Sanitization: Absent. The skill relies on human confirmation of the drafted MARKETING.md file rather than automated filtering.
- [DATA_EXPOSURE]: The skill's discovery phase searches for email, advertising, and analytics configurations to build product context. While intended for marketing analysis, this could potentially surface sensitive identifiers or metadata if these configurations contain private keys or account secrets.
Audit Metadata