feedback-synthesis
Pass
Audited by Gen Agent Trust Hub on Jul 31, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill is designed to ingest and process 'user feedback' from various external and internal sources (Support tickets, App reviews, User interviews, Surveys, Social media). This constitutes an untrusted data ingestion surface where malicious actors could embed instructions intended to influence the agent's behavior during the synthesis process.
- Ingestion points: Processes data from sources identified in SKILL.md under 'Feedback Sources' and 'Context Integration' (e.g.,
docs/product/feedback-reports/[period].md). - Boundary markers: The skill lacks explicit instructions or boundary markers to treat the ingested feedback strictly as data rather than instructions.
- Capability inventory: The skill performs file read/write operations (
docs/product/.ompm/feedback-synthesis.json) and triggers subsequent skills (retrospective,product-strategy). A successful injection could theoretically corrupt these reports or influence prioritization in the product strategy. - Sanitization: There are no documented sanitization, filtering, or escaping mechanisms for the raw user quotes being processed and displayed in the output.
Audit Metadata