skill-quality-gate · Pre-Publish Quality Gate
A skill for the final check before publishing: one review, two gates — compliance risk detection plus deliverable quality review. Compliance covers absolute claims, sensitive wording, regulated categories, and platform-specific restrictions; quality covers completeness, readability, and platform fit.
Submit text, an image path, or a video path, optionally naming the target platform to load its rule set. The verdict: ✅ ready / ⚠️ needs fixes / ❌ not up to standard, with the top 3 fixes attached.
Example invocation: "Review this Xiaohongshu post — is it safe to publish?" with the copy attached.
Full brief
Positioning
skill-quality-gate is the last gate before publishing: one skill, two checks — compliance risk detection and deliverable quality review. It judges whether content is safe to ship and what needs fixing; it generates no new content.
Core capabilities
- Compliance detection: item-by-item checks for absolute claims, regulated-category violations, prohibited content, and platform-specific restrictions — general rules plus platform-specific rule sets (Xiaohongshu, Douyin, Bilibili each have their own; generic rules only when no platform is specified).
- Quality review: per-dimension review by artifact type (text/image/video) — completeness, readability, platform fit.
- Three-tier verdict: ✅ pass, publish directly / ⚠️ flaws, minor polish then publish / ❌ below standard, rework — with rework guidance when it fails.
- Composite rules: high compliance risk or quality rework → overall ❌; low compliance risk plus quality pass → overall ✅; any other combination → ⚠️ needs fixes.
- Top 3 fixes: every run ships the three highest-priority fixes.
Workflow
- Gate 1 — compliance: read content → load general and platform rules → item-by-item checks → compliance summary
- Gate 2 — quality: identify artifact type → per-dimension review → three-tier verdict → rework guidance if failed
- Composite verdict: apply the combination rules, output the final verdict and top 3 fixes
Inputs & outputs
| Input | Notes |
|---|---|
| Content to review | Required — text, image path, video path, or mixed |
| Target platform | Optional; loads that platform's rule set when specified |
Output: a JSON report — overall verdict, compliance detail (risk level / issue list / passed checks), quality review (dimension scores / tier), and the top 3 fixes.
Boundaries with adjacent skills
- skill-post-scorer: scores a draft's engagement potential (how good); this skill checks publishability (compliant, up to standard). They chain naturally: score and polish first, then pass the gate.
Fit
- Dual compliance-and-quality check before every publish
- Pre-review of copy with sensitive claims (efficacy, price comparisons)
- A shared publish bar for teams, beyond any single person's judgment
Before you start
- Provide the full content or file path; naming a target platform loads its dedicated rules.
- With a profile, the review adapts to the profile's platform and style automatically.
- No API key or fees involved.
skill-quality-gate is part of the Aiglade Skill library. Invoke it from the Aiglade chat box in plain language.