Mandatory workflow for ALL content production — blog articles, social posts, email copy, ad copy, release notes. Every piece of content MUST go through produce → review → revise. Auto-spawn a review agent after every content draft. This skill triggers whenever you write, draft, or produce any content.
SkillScore55Maintained
Automatic
◆Marketing & Content pack
Last verified 10/7/2026
What it changes
Passed 12 of 22 test cases with the skill, and 4 of 22 without it.
Without skill18%
With skill55%
Checks it passed that the model alone missed
Avoid inflating word count with generic filler.
Explicitly label the review pass as a self-review.
Refuse to inflate FAQ items to meet template targets.
and 7 more in the full report
8 fixed4 passed either way10 still failing0 broke
Pass-rate lift
+36 pp
Tokens
+13%
Agent turns
0%
Response time
+2%
Model gemini-3.6-flashMeasured 10/7/2026v312 of 22 cases passed with skill
Tokens, turns and time describe resource use, not a dollar price.
See it in action
One of 8 test cases the skill fixed. Same model, same prompt, with and without the skill.
1 of 8
Prompt
Our article '2025 Enterprise Security Trends' was published yesterday at https://example.com/blog/2025-security-trends. Perform a post-publication technical audit on this live URL.
What was checked, without and with the skill
Checked
Without skill
With skill
Confirm post-publication audit coverage applies to the live URL.
Missed
Passed
case-16 · gemini-3.6-flash · 10/7/2026 · answers are capped and redacted
What's inside
The instructions your model reads when the skill loads.
SKILL.md11 linessingle file
Content production
Match the requested deliverable, language, audience and existing approval checkpoints. A full article task includes a finished draft and relevant checks; a topic, edit or review request stops at that requested output. Do not create extra approval checkpoints or publish merely because you produced a package.
Read supplied context and use the installed skill matching the task: seo-blog-writer for site SEO research/drafting, writer for general articles, a platform writer for channel conventions, or justeng-wechat for JustEng. These are separately installed capabilities, not files assumed inside this package. If unavailable, complete the applicable work directly and state the actual coverage.
Research material claims; preserve source context, author evidence and uncertainty. Write in the audience’s language. Use the matching bilingual localizer when translating supplied text.
Review evidence, reader value, whole-piece voice and destination constraints. content-qa can provide a separate editorial review; seo-geo-qa provides technical checks when applicable. Independent delegation is optional when supported and authorized; do not require a named model or host-specific spawn API. Without a separate reviewer, label the pass as self-review.
Fix demonstrated issues, then recheck changes and outstanding findings. Do not inflate link counts, FAQ, word counts or keyword density to satisfy default templates.
Deliver the requested files and relevant evidence/QA notes. Distinguish technical results, editorial review, human approval and actual publication. Public uploads and publishing use existing authorization for those actions.
After publication, an installed seo-geo-qa can audit the supplied live URL when requested. It is a separate task with its own dependency and coverage report; do not claim it was executed on an unpublished draft.
Generate an image with ChatGPT (GPT image generation) in the user's logged-in ego lite browser and save the file into the current project directory. Use this skill whenever the user wants an AI-generated picture — 生成图片、画一张图、做海报/攻略图/插画/封面/信息图, "generate an image of...", "make a poster/infographic/illustration" — especially when they give an image description and want the result saved locally for project use. Prefer this skill whenever ego-browser + ChatGPT image generation is the intended pipelin