orchestration-patterns

Pass

Audited by Gen Agent Trust Hub on Jul 30, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is designed to ingest and act upon data from external, potentially untrusted sources such as issue trackers and the project codebase.
  • Ingestion points: The agent reads acceptance criteria from tickets (e.g., {{TICKET_PREFIX}}-YYY) and searches the codebase and previous sessions during the discovery phase.
  • Boundary markers: There are no instructions for the agent to use delimiters or specific safety instructions to disregard embedded commands when processing this external data.
  • Capability inventory: The skill empowers the agent to execute shell commands (via {{CI_VALIDATE_COMMAND}}), write to the file system (e.g., docs/agent-outputs/qa-validations/), and interact with external ticket systems.
  • Sanitization: No sanitization, escaping, or validation of the ingested external content is mentioned before it is used to influence the agent's logic or execution loop.
  • [COMMAND_EXECUTION]: The orchestration loop involves executing shell commands provided through the {{CI_VALIDATE_COMMAND}} placeholder. While standard for development workflows, this mechanism allows the agent to execute arbitrary code within the execution environment based on project configuration.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 30, 2026, 04:40 PM
Security Audit — agent-trust-hub — orchestration-patterns