kai-growth-plan
Pass
Audited by Gen Agent Trust Hub on Nov 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from the project being analyzed, including README.md, package.json, landing pages, and other project files, to generate the MARKETING.md file and growth plan. An attacker could potentially embed malicious instructions within these project files to influence the agent's output.
- Ingestion points: MARKETING.md, CLAUDE.md, README.md, PROJECT.md, package.json, and general codebase exploration including landing pages.
- Boundary markers: None present to delimit or ignore instructions within ingested content.
- Capability inventory: Reads project files and writes markdown files to the workspace/growth-plan/ directory.
- Sanitization: No explicit sanitization or instruction-ignoring delimiters are defined for the ingested content.
- [DATA_EXFILTRATION]: The instructions direct the agent to search for "email/ad/analytics config" files during project exploration. While intended to identify marketing tools for the audit, this pattern directs the agent to locate and potentially process sensitive configuration files that might contain API keys or service credentials if stored within the project structure.
Audit Metadata