The Agent Skills Directory

[SAFE]: The skill implements a robust credential resolution process in scripts/gen.py to allow authenticated access to image generation services.
It scans for OPENAI_API_KEY in environment variables and local .env files.
It accesses platform-specific authentication tokens at ~/.pi/agent/auth.json to facilitate integration with the host agent.
This behavior is clearly documented in SKILL.md and references/auth-pitfalls.md.
[SAFE]: The skill performs dynamic path manipulation to access internal agent libraries.
scripts/gen.py modifies sys.path to include the agent's installation directory for importing agent.auxiliary_client.
This is used to read authorized Codex tokens from the host platform's internal store.
[SAFE]: All network activity is directed toward well-known and trusted endpoints.
Data such as prompts and images are transmitted to api.openai.com and chatgpt.com.
The urllib.request module is used to retrieve generated images from URLs provided by the official OpenAI API.

gpt-image-gen