skills/q00/openclip/oc-stt-worker/Gen Agent Trust Hub

oc-stt-worker

Pass

Audited by Gen Agent Trust Hub on Sep 5, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes the Bash tool to execute a project-specific CLI command (oc) for audio transcription. This execution is scoped to specific project and chunk parameters.
  • [INDIRECT_PROMPT_INJECTION]: The skill acts as a data processor that transcribes external audio chunks. This creates an indirect prompt injection surface where malicious instructions could be embedded in the audio source and processed into text for downstream components.
  • Ingestion points: Audio data chunks processed via the oc stt command (SKILL.md).
  • Boundary markers: None are defined to isolate the transcribed content from agent instructions.
  • Capability inventory: The skill uses Bash for command execution and Read to verify the presence of segment files (SKILL.md).
  • Sanitization: There are no documented steps for sanitizing or escaping the transcribed output before it is returned to the orchestrator.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 5, 2026, 10:33 PM
Security Audit — agent-trust-hub — oc-stt-worker