sdd-design
Pass
Audited by Gen Agent Trust Hub on Mar 14, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: Indirect prompt injection surface detected.
- Ingestion points: The skill retrieves proposal and specification content from memory via
mem_get_observationand performs a comprehensive scan of the project codebase in Step 2. - Boundary markers: Absent. The instructions do not define delimiters or provide specific warnings to the agent to disregard instructions potentially embedded within the source documents.
- Capability inventory: The skill has the capability to write files to the local filesystem (
design.md) and update the persistent memory store via themem_savetool. - Sanitization: No sanitization, escaping, or validation of the ingested external content is performed before it is used to generate output.
Audit Metadata