openai-symphony-autonomous-agents

Pass

Audited by Gen Agent Trust Hub on Mar 17, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: Fetches the reference implementation and specification from OpenAI's official GitHub repository.
  • [COMMAND_EXECUTION]: Includes standard commands to manage the Elixir project lifecycle, including dependency installation and service startup.
  • [PROMPT_INJECTION]: The skill ingests external task data from Linear to drive the autonomous coding agent. This creates a surface for indirect prompt injection if malicious instructions are embedded in task descriptions.
  • Ingestion points: Task titles and descriptions are retrieved from the Linear GraphQL API.
  • Boundary markers: The code snippets do not illustrate explicit boundary delimiters for the agent context.
  • Capability inventory: The runner interacts with git, the Codex API, and GitHub to automate code changes and PR creation.
  • Sanitization: Includes basic string normalization for branch names.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 17, 2026, 06:58 AM