create-speech-recording-controls

Pass

Audited by Gen Agent Trust Hub on Oct 6, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes text provided by the user or generated by the agent and passes it to a shell command, creating a potential vector for indirect prompt injection.
  • Ingestion points: Untrusted data enters via the $RESPONSE_AS_TEXT variable and the $RESPONSE_AS_TEXT_FILE reference in SKILL.md.
  • Boundary markers: The instructions do not specify explicit delimiters or warnings to ignore embedded instructions within the text being processed.
  • Capability inventory: The skill executes subprocesses via the agent-voice speak command in SKILL.md.
  • Sanitization: The instructions include basic normalization by converting literal control characters (e.g., \n) into spoken forms.
  • [EXTERNAL_DOWNLOADS]: The skill involves the installation of a CLI tool from a remote source.
  • Evidence: The Setup section in SKILL.md instructs the user to run uv tool install agent-voice and agent-voice controls install.
  • Source: The tool is sourced from the author's official GitHub repository mentioned in the Resources section.
  • [COMMAND_EXECUTION]: The skill relies on shell command execution for its primary functionality.
  • Evidence: Multiple shell command blocks in SKILL.md demonstrate the use of agent-voice to record and play back speech.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 6, 2026, 11:04 AM