The Agent Skills Directory

[SAFE]: The skill operates as a quality assurance gate, using standard diagnostic tools to verify local media artifacts. No external network requests or sensitive data access patterns were identified.
[COMMAND_EXECUTION]: The skill utilizes the 'Bash' tool to execute 'ffmpeg' and a project-specific 'oc' utility. These executions are restricted to diagnostic verification of deliverables and do not interact with sensitive system files or administrative functions.
[PROMPT_INJECTION]: The instructions establish a robust adversarial verification posture, directing the agent to doubt the validity of artifacts until evidence is confirmed. This internal logic does not attempt to bypass platform safety guidelines or override agent behavior maliciously.
[SAFE]: (Indirect Prompt Injection Surface Check)
Ingestion points: Processes media files and SRT transcripts provided as input to the verification commands.
Boundary markers: The skill relies on external CLI tool outputs rather than direct LLM parsing of raw binary or text data for logic control.
Capability inventory: Limited to 'Bash' for executing specific verification commands ('oc verify', 'ffmpeg').
Sanitization: The use of specialized verification binaries (oc, ffmpeg) provides a layer of isolation between the deliverable content and the agent's reasoning instructions.

oc-verifier