The Agent Skills Directory

[SAFE]: The skill interacts with the Volcengine TTS API at openspeech.bytedance.com, which is a well-known service provider. This network operation is directly related to the skill's primary purpose of speech synthesis.
[SAFE]: The skill manages authentication using environment variables (VOLCENGINE_TTS_APPID and VOLCENGINE_TTS_ACCESS_TOKEN), which is the recommended method for handling sensitive API keys and prevents credential exposure in the source code.
[SAFE]: The Python script scripts/generate.py performs standard operations including JSON parsing, multi-threaded API requests, and file I/O within designated directories. It does not employ dangerous execution patterns or attempt to escalate privileges.
[PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface because it processes untrusted user data to generate podcast dialogue. Ingestion points: User-provided source content entered during the first step of the workflow in SKILL.md. Boundary markers: There are no specific delimiters or instructions to ignore embedded commands within the processed text. Capability inventory: The skill can write JSON, MP3, and Markdown files to the workspace and make network calls to the TTS API. Sanitization: No sanitization or filtering of the user-provided text is performed before it is included in the script JSON. Evidence Note: While the surface exists, the risk is minimal as the content is synthesized into audio rather than being used to construct executable commands.

podcast-generation