Source Provenance Guard

SkillProductivity

Verifies that a ClipTalk Agent plan uses the current task, source revision, selected scope, and task-local artifacts before reusing analysis, covers, subtitles, timelines, or outputs.

Available today. Use it from your connected AI after setup.

Connect ahel once, and every AI you use reads what you have installed.

Then ask your AI: use the Source Provenance Guard skill

What this skill tells your AI

The instructions your AI receives, as published by gml-mmgroup/cliptalk in skills/cliptalk-source-provenance-guard/SKILL.md and read by ahel’s review.

Prevent cross-task state leakage before an edit or delivery step.

  • Inspect the active workspace and validate the bound job, source asset, source scope, active plan, edit session, cover, subtitle draft, and output references.
  • Treat a same-source upload as a new task unless the user explicitly asks to continue the previous task.
  • Never reuse a previous task's time range, cover, subtitle draft, candidate selection, timeline, render, or progress state as the current task's result.
  • Return a clear prerequisite or mismatch diagnosis instead of repairing by silently switching jobs.
  • This Skill does not render or export media.

Signals

GitHub stars
126
Forks
8
Last commit
Sep 2026
Advanced
Catalog kind
skill
Gateway key
cliptalk-source-provenance-guard
Source
github.com/gml-mmgroup/cliptalk