dt-app-dashboards

Pass

Audited by Gen Agent Trust Hub on Sep 14, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes dashboard JSON documents retrieved from an external server, which introduces a risk of indirect prompt injection if the content contains malicious DQL queries or instructions.
  • Ingestion points: Dashboard content is fetched via dtctl get dashboard as instructed in SKILL.md and references/create-update.md.
  • Boundary markers: The skill provides an example JSON structure in SKILL.md and instructions to inspect available properties.
  • Capability inventory: The skill utilizes dtctl query to execute DQL, dtctl apply to update server-side resources, and performs local file system operations (write/read/delete) for dashboard.json.
  • Sanitization: references/create-update.md specifies a mandatory DQL validation process using dtctl query with the --plain flag and advises using single quotes for shell command safety.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 14, 2026, 05:22 PM
Security Audit — agent-trust-hub — dt-app-dashboards