The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes local Python scripts (scripts/processon_api_client.py and scripts/check_update.py) to interact with the ProcessOn API and perform maintenance tasks. This is standard functionality for the skill's purpose.
[EXTERNAL_DOWNLOADS]: The skill fetches version information from the vendor's official GitHub repository (processonai/processon-skills). This is a safe operation used exclusively for checking for skill updates.
[CREDENTIALS_UNSAFE]: The skill manages a sensitive API key (PROCESSON_API_KEY) using environment variables. It includes explicit instructions to the agent to avoid leaking the key in outputs and employs standard Authorization headers for all API requests.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied prompts to generate diagram specifications, creating a potential ingestion surface for indirect prompt injection.
Ingestion points: User-provided prompt and title arguments in scripts/processon_api_client.py.
Boundary markers: Absent.
Capability inventory: The skill has network access via urllib.request and file system write access for saving rendered images in scripts/processon_api_client.py.
Sanitization: Diagram titles are sanitized using a regex-based slugification function to prevent path traversal, and all API payloads are strictly JSON-encoded.
[OBFUSCATION]: The skill uses Base64 decoding specifically to process image data returned by the ProcessOn API, which is an expected and safe implementation for image handling.

processon-diagram-generator