sla-breach-escalation

Pass

Audited by Gen Agent Trust Hub on Apr 12, 2026

Risk Level: SAFENO_CODE
Full Analysis
  • [SAFE]: The skill consists entirely of markdown instructions and templates for helpdesk operations.
  • [NO_CODE]: No scripts, executables, or code snippets were found in the skill content. It relies on standard template interpolation.
  • [DATA_EXPOSURE]: The skill utilizes configuration fields for ticket metadata (Ticket ID, Priority, SLA details). These are used solely for generating reports within the agent context.
  • [PROMPT_INJECTION]: No override markers, role-play attempts, or malicious instructions were detected in the text or metadata.
  • [INDIRECT_PROMPT_INJECTION]: The skill defines a surface for untrusted data ingestion.
  • Ingestion points: Configuration fields ticket_id, sla_type, sla_target, current_elapsed, and priority from the skill frontmatter are interpolated into the body.
  • Boundary markers: Absent; data is placed directly into markdown headers and text blocks.
  • Capability inventory: No access to subprocess calls, file-system operations, or network tools was identified in the skill configuration or content.
  • Sanitization: None specified for the input variables.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 12, 2026, 04:21 PM
Security Audit — agent-trust-hub — sla-breach-escalation