YouTube Transcript Pipeline Lite

Run a lightweight YouTube transcript workflow: transcribe, attribution cleanup, translation, and packaging with minimal tooling. Use for repeatable transcrip...

This skill is instruction-only and internally consistent, but it depends on whatever transcription tooling the agent already has: before installing, confirm the agent's transcript pipeline (or any external APIs it will invoke) is trustworthy and won't leak sensitive audio/transcript data. Because the skill keeps raw outputs and manifests, make sure you have an appropriate data-retention and sharing policy for transcripts and translations. Limit use to explicit invocations (don't bind it to always-on automation) if you want tighter control.

Type: OpenClaw Skill Name: youtube-transcript-pipeline-lite Version: 1.0.0 The skill bundle describes a legitimate workflow for processing YouTube transcripts, including cleanup, translation, and packaging. All instructions in `SKILL.md` and `references/limits.md` are aligned with the stated purpose and include guardrails for the AI agent's behavior. There is no evidence of prompt injection attempts, data exfiltration, malicious execution, persistence mechanisms, or obfuscation. The implied file system operations (creating directories/files) and the instruction to 'Run your existing transcript flow' are within the expected capabilities of an AI agent for this type of task and do not indicate malicious intent.