discover-interview-synthesis
Pass
Audited by Gen Agent Trust Hub on Oct 17, 2026
Risk Level: SAFEPERSISTENCEINDIRECT_PROMPT_INJECTION
Full Analysis
- [PERSISTENCE]: The skill utilizes a "Project Memory Contract" to maintain state across sessions. It instructs the agent to read from and write findings to a local file (
.claude/pm-skills.local.md), which allows it to share context (like personas and interpretations) with other skills in the product-on-purpose suite. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user research data, which is a known vector for indirect prompt injection where malicious instructions could be embedded in transcripts.
- Ingestion points: The skill's core workflow involves gathering raw interview notes, transcripts, and recordings from external sources (SKILL.md, Step 1).
- Boundary markers: There are no defined delimiters or instructions to the agent to disregard embedded commands or formatting within the source research materials.
- Capability inventory: The skill has the capability to write research summaries and synthesized data to the local project memory file.
- Sanitization: No sanitization or filtering steps are specified for the input text before it is processed and summarized by the agent.
Audit Metadata