call-analysis

Warn

Audited by Gen Agent Trust Hub on Feb 16, 2026

Risk Level: MEDIUMPROMPT_INJECTION
Full Analysis
  • Prompt Injection (MEDIUM): The skill is vulnerable to indirect prompt injection (Category 8) due to its processing of external, untrusted content.
  • Ingestion points: The skill ingest 'sales call transcripts' which are external and potentially attacker-controlled (e.g., a malicious participant on a call).
  • Boundary markers: There are no specified delimiters or boundary instructions provided in the SKILL.md to help the agent distinguish between the transcript text and the analysis instructions.
  • Capability inventory: While the skill itself is a text analyzer, its output is explicitly cross-referenced to drive tools with external side effects, such as 'follow-up-emails', 'multithread-outreach', and 'account-qualification'.
  • Sanitization: The instructions contain no sanitization or validation steps to filter out potential system-override commands embedded within transcripts.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 16, 2026, 07:52 AM