anti-ai-slop
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted project source code (HTML, CSS, JavaScript) to perform design audits and identity extraction. This ingestion provides a potential surface for indirect prompt injection if those files contain malicious instructions designed to subvert the agent's behavior.
- Ingestion points: Untrusted source code is read into the context both through the main agent workflow and via the output of the
scripts/extract-identity.mjsscript. - Boundary markers: The skill does not explicitly define delimiters or boundary instructions to isolate product content from the agent's core logic.
- Capability inventory: The agent is granted capabilities to read and write files within the project directory to implement design repairs.
- Sanitization: No sanitization or filtering for natural language instructions is performed on the product source code during the extraction or auditing process.
Audit Metadata