The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the Bash tool to execute oc CLI commands for mechanical verification and ffmpeg for video sampling and analysis.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data, including subtitle files (.srt), transcripts, and project manifests, to verify deliverable quality. This represents an attack surface where instructions embedded in data could influence agent behavior.
Ingestion points: Ingests oc output JSON, .srt files, and transcript text during verification steps.
Boundary markers: The instructions do not define clear delimiters or "ignore instructions" prompts when reading subtitle or transcript content.
Capability inventory: Uses Bash (executing oc and ffmpeg) and Read (accessing the file system) to perform its checks.
Sanitization: There is no evidence of sanitization or filtering applied to the content of the SRT or transcript files before processing.

oc-verifier