The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses standard CLI tools curl and jq to execute API requests and process JSON responses.
[EXTERNAL_DOWNLOADS]: Interacts with api.together.ai, which is the official and well-known endpoint for Together AI services. It also references vendor-specific resources at app.vm0.ai.
[DATA_EXFILTRATION]: Employs the $TOGETHER_TOKEN environment variable for secret management, which is a recommended security practice to avoid hardcoding credentials.
[INDIRECT_PROMPT_INJECTION]:
Ingestion points: User-supplied text is used as content for the messages, prompt, and input fields in JSON payloads written to /tmp/.
Boundary markers: No explicit delimiters are present in the JSON structures to isolate untrusted user data.
Capability inventory: The skill allows network communication to the Together AI API and file writing to the /tmp/ directory.
Sanitization: The skill does not perform explicit validation or escaping of the user-provided prompts before they are sent to the external API.

together