skill-risk-scanner · Originality & Copyright Risk Assessment
Assess whether content shows rewriting/plagiarism patterns, media copyright risk, and citation compliance — delivered as a qualitative risk report.
LLM-text-analysis based: no technical plagiarism detection, no similarity scores; only low/medium/high qualitative grades. Image/audio risk is judged from user-provided source descriptions only — never guessed. Risks split into "must fix" and "advisory"; every recommendation names exactly which paragraph to change or which asset to replace.
Example invocation: "Check this draft for rewriting-plagiarism risk."
Full brief
Positioning
skill-risk-scanner is the pre-publish content risk check: it answers "is this safe to publish". It covers originality and copyright-infringement risk — not platform rule compliance.
Core capabilities
- Rewriting-pattern scan: detects mechanical synonym swaps, rephrased-but-same-argument structures, and rearranged paragraph moves; each finding quotes the specific passage with reasoning.
- Uniqueness assessment: checks for personal experience / exclusive data / original viewpoints / concrete details; concludes high / average / low originality.
- Citation compliance: verifies source attribution for opinions, data, screenshots; when a reference text is provided, compares paraphrase thoroughness.
- Media copyright risk: judged solely on user-provided source descriptions (images/audio); marked "cannot assess" when absent — never guessed; high-risk assets get free replacement suggestions.
- Brand/trademark risk: scans for competitor disparagement, unauthorized brand-name use, false endorsement implications.
- Risk report: overall low/medium/high grade + "must fix"/"advisory" lists + actionable recommendations.
Workflow
- Confirm inputs: draft, source descriptions, reference text, target platform
- Rewriting scan (per paragraph) → uniqueness assessment → citation check
- Media copyright assessment (only when sources described) → brand/trademark scan
- Assemble report, split must-fix vs advisory
Inputs & outputs
| Input | Required | Notes |
|---|---|---|
| Draft to check | Yes | Copy, script, or long-form draft |
| Asset source descriptions | Optional | Image/audio/video source notes |
| Reference text | Optional | Used to judge rewrite thoroughness |
| Target platform | Optional | Matches platform-specific copyright rules |
Output: Markdown risk report (overall grade + originality + media risk + brand risk + risk lists + recommendations).
Boundaries with adjacent skills
| Skill | Lane |
|---|---|
| skill-risk-scanner (this) | Originality + copyright-infringement risk |
| skill-quality-gate | Platform compliance + content quality (sensitive words / ad law / platform rules) |
| skill-publish-checklist | Completeness checks (title/cover present?) |
Fit
- Pre-publish originality self-checks
- Rewriting/plagiarism suspicion screening
- Media copyright pre-screening
- Brand-collaboration infringement review
Before you start
- No API key; this is LLM text analysis, not technical plagiarism detection.
- Boundaries: no similarity percentages or originality scores; no image-content recognition or audio fingerprinting; undescribed sources can't be assessed.
- The report honestly states its limits: use Copyleaks, Originality.ai, or similar professional tools for precise plagiarism checks.
skill-risk-scanner is part of the Aiglade Skill library. Invoke it from the Aiglade chat box in plain language.