get-rep-call-feedback
Pass
Audited by Gen Agent Trust Hub on Jul 15, 2026
Risk Level: SAFE
Full Analysis
- Data Ingestion Analysis: The skill is designed to process external meeting transcripts and user-provided files. While ingesting external data is a common vector for indirect prompt injection, the skill incorporates behavioral guidelines to focus on short, specific excerpts, which helps maintain the context and reduces the risk associated with processing untrusted data.
- Link and Identifier Handling: The instructions specifically mandate using readable context for citations and explicitly forbid exposing internal connector IDs in the output. This practice supports better security by preventing the exposure of underlying system identifiers.
- Controlled Workflow: The skill defines clear boundaries for its operations, focusing solely on synthesis and coaching readout without attempting to perform automated actions like sending messages or creating formal scorecards, which limits the scope of its automated capabilities to a safe environment.
Audit Metadata