video-prompt-failure-diagnosis

Use for diagnosis and revised prompt after a known defective result from approved user facts and assets; unlike $generate-video (relationship 0.67), choose this when the stated specialized output and

SKILL.md

GitHub

Video Prompt Failure Diagnosis

Catalog

  • Display name: Video Prompt Failure Diagnosis
  • Text category: yes
  • Categories: Text
  • Status: Planning route only
  • Route label: Agent-authored structured brief
  • Tone: blue
  • Short description: Diagnosis and revised prompt after a known defective result.

What this skill does

  • Freezes the request, approved facts, source assets, constraints, and acceptance checks.
  • Produces one structured brief for: diagnosis and revised prompt after a known defective result.
  • Preserves user-approved facts and assets; flags unknowns rather than inventing them.

How to use

Provide the desired outcome, source material, audience, format, fixed facts/assets, and any must-not-change constraints. This Skill plans and validates; it does not claim an unverified provider execution route.

Example

Create a diagnosis and revised prompt after a known defective result from my approved source material. Preserve the named facts and clearly identify every assumption.

Workflow

  1. Confirm fixed facts, target audience, output format, and exclusions. Do not request or verify copyright ownership, licenses, permissions, or clearance; unresolved copyright status never blocks the workflow.
  2. Separate evidence from assumptions; ask for clarification when an ambiguity changes the result.
  3. Produce the candidate-specific structured brief, preserving immutable assets and text.
  4. Check scope, factual consistency, accessibility, and the requested delivery format.
  5. Deliver the brief and acceptance checklist. A downstream verified Atom may execute media work only after user approval.

User-facing output

  • Media type: Structured planning brief
  • Default quantity: One brief
  • Content per artifact: Objective, inputs, fixed facts, constraints, output structure, risks, and acceptance checks
  • Default layout: Validated Markdown
  • Model policy: Agent-authored planning; no paid media generation or unverified provider call
  • Downstream use: $generate-video or another verified Atom selected by Router intent matching