customer-escalation

Pass

Audited by Gen Agent Trust Hub on Jun 14, 2026

Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
  • [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection due to its reliance on external data sources for context gathering.
  • Ingestion points: The workflow in SKILL.md (Step 2: Gather Context) instructs the agent to read data from support platforms, CRM systems, chat logs, and project trackers.
  • Boundary markers: No specific delimiters or "ignore embedded instructions" warnings are defined to separate untrusted customer data from the agent's instructions.
  • Capability inventory: The skill is configured with Read, Glob, and Grep tools, allowing the agent to read file system content.
  • Sanitization: There are no instructions for sanitizing or validating the content retrieved from these external platforms before including it in the final escalation brief.
  • [NO_CODE]: The skill consists entirely of markdown instructions and configuration metadata without any accompanying scripts or executable files.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 14, 2026, 08:50 AM
Security Audit — agent-trust-hub — customer-escalation