The Agent Skills Directory

[SAFE]: The skill provides legitimate functionality for interacting with the StepFun TTS API. All bundled scripts and documentation are consistent with the stated purpose of generating and comparing speech synthesis outputs.\n- [COMMAND_EXECUTION]: The utility script scripts/ab_compare.sh uses standard system tools such as ffprobe, stat, and awk to compare audio file properties. The synthesis script scripts/tts_generate.py uses standard networking libraries. No signs of command injection or malicious system modification were found.\n- [CREDENTIALS_UNSAFE]: Secret management is handled correctly using environment variables or a local configuration file (config.json). The skill does not contain hardcoded API keys or sensitive credentials.\n- [INDIRECT_PROMPT_INJECTION]: The skill documents an ingestion surface for prosody directives embedded in synthesized text.\n
Ingestion points: Untrusted text is processed via the --text and --batch arguments in scripts/tts_generate.py.\n
Boundary markers: The system uses () parentheses to delimit instructions meant for the TTS model.\n
Capability inventory: The skill includes network communication via urllib.request and local file writing for audio output.\n
Sanitization: There is no sanitization of the input text, as the prosody directives are a desired feature of the underlying service.\n
Context: This is a documented operational feature of the StepFun service and does not present a threat to the agent's environment.

stepfun-tts