support-feedback-prioritization
Pass
Audited by Gen Agent Trust Hub on Oct 10, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external customer-facing sources (Intercom and Slack) to categorize and score feedback.
- Ingestion points: The skill reads all tickets/conversations from Intercom and specified Slack channels (SKILL.md, Step 1).
- Boundary markers: The prompt template lacks explicit delimiters or instructions to treat ingested signal text as data only, increasing the risk that an LLM might follow instructions embedded within a customer ticket.
- Capability inventory: The skill has the ability to write files to Google Drive (SKILL.md, Step 5).
- Sanitization: There is no mechanism described to sanitize or filter the raw feedback text before it is processed for scoring and reporting.
Audit Metadata