execute-ticket-graph

Pass

Audited by Gen Agent Trust Hub on Aug 16, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill processes untrusted content (titles, bodies, and comments) from external trackers like GitHub, GitLab, and Jira to drive subagent tasks, creating a surface for indirect prompt injection.\n
  • Ingestion points: Ticket data is fetched from external trackers during the discovery and normalization phase described in SKILL.md.\n
  • Boundary markers: While data is passed in a structured packet to subagents, the instructions lack explicit directives for subagents to ignore embedded instructions in ticket data.\n
  • Capability inventory: Capabilities include subagent orchestration, Git worktree/branch management, and execution of verification commands.\n
  • Sanitization: Input is normalized into a JSON snapshot and validated for structural integrity by a bundled script, but natural language content is not sanitized.\n- [COMMAND_EXECUTION]: The skill runs a local Node.js script scripts/inspect-graph.mjs to validate the ticket dependency graph. This script performs structural checks and simulations based on local JSON snapshots.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 16, 2026, 08:16 PM
Security Audit — agent-trust-hub — execute-ticket-graph