audio-transcription
Pass
Audited by Gen Agent Trust Hub on Sep 21, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill relies on the
postplusCLI for its primary functions, includingpostplus media transcribe,postplus media poll,postplus doctor, andpostplus update. This gives the skill significant interaction with the shell environment.\n- [INDIRECT_PROMPT_INJECTION]: The skill ingests audio data from external sources and transforms it into text that is then integrated into the agent's conversational context.\n - Ingestion points: The skill instructions allow for the
--audioflag to accept HTTPS URLs and local paths (SKILL.md).\n - Boundary markers: The skill does not provide delimiters or specific instructions to the agent to distinguish between its primary instructions and the content of the generated transcript.\n
- Capability inventory: The skill has the capability to execute shell commands using the
postplustool and perform file system operations within the.postplus/directory.\n - Sanitization: There is no mechanism described for sanitizing the resulting transcription to prevent potential prompt injection attacks hidden within the audio content.
Audit Metadata