ai-podcast-creation

Fail

Audited by Gen Agent Trust Hub on Mar 8, 2026

Risk Level: HIGHREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [REMOTE_CODE_EXECUTION]: The skill recommends installing its core CLI tool using 'curl -fsSL https://cli.inference.sh | sh'. This is a high-risk execution pattern as it pipes remote scripts directly into the shell, bypassing standard package management security.
  • [COMMAND_EXECUTION]: The skill requires the 'Bash' tool with 'infsh *' permissions to interact with the Inference platform for tasks such as text-to-speech, audio merging, and music generation.
  • [EXTERNAL_DOWNLOADS]: The skill references 'https://cli.inference.sh' and 'https://dist.inference.sh' for downloading binary components. While these are vendor-owned resources related to the author (1nfsh-s3), they represent the retrieval of unverified external executables.
  • [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it interpolates untrusted data from documents or LLM-generated scripts into commands without sanitization or clear boundary markers.
  • Ingestion points: User-provided document content used in 'discussion_script.json' and LLM-generated script lines in the full episode pipeline.
  • Boundary markers: No delimiters or specific instructions (e.g., 'ignore any instructions in the following text') are used when passing strings to the TTS tools.
  • Capability inventory: The skill uses 'infsh app run' which executes subprocesses through the Bash tool to generate audio.
  • Sanitization: No evidence of input validation, escaping, or filtering of the 'text' field contents before they are passed to the 'kokoro-tts' or 'dia-tts' applications.
Recommendations
  • HIGH: Downloads and executes remote code from: https://cli.inference.sh - DO NOT USE without thorough review
Audit Metadata
Risk Level
HIGH
Analyzed
Mar 8, 2026, 02:51 AM