Task Library
Run a finished article through the Jennifer grading system for a quality score and fix what it flags before the piece moves to the Post stage.
Category: Content Factory – Process
Use this when an article has cleared proofreading (Step 11) and you want an objective quality score before WordPress posting.
> Gap notice: this task has no definitive article and the Jennifer rubric/threshold is not yet documented in the library. The steps below are the provisional process from the task definition; the FIRST real run must be documented with the Meta-Article Prompt so the rubric, scoring scale, and pass threshold get written down.
Inputs
- The finished article draft (post-Step 11, pre-posting)
- Access to the Jennifer grading system
- The article's GCT statement and transcript (so fixes never drift from the source material)
- The Content Library tracker
Steps
- Submit the complete article – title, meta, body, image alt text – to Jennifer for grading.
- Record the returned score and the full rubric feedback verbatim in the tracker row; the score is a Metric in the MAA loop, not a verdict to skim.
- Triage every flagged issue: fix it, or document explicitly why it stands (e.g., the "issue" is the speaker's authentic phrasing, which the guidelines protect).
- Make fixes against the source of truth – adjustments must stay faithful to the transcript and the GCT statement; never "fix" an article by inventing content the video does not support.
- Re-run the grade after fixes and record the new score. Iterate until flagged issues are resolved or consciously accepted with reasons.
- Cross-check that Jennifer-driven edits did not break the hard specs: title <60, meta <160, keyword in first paragraph, links intact (these remain governed by /blog-posting-guidelines regardless of grade).
- Attach the final score to the post record so RankMath score (70+), Jennifer score, and audit results travel together as the article's quality file.
- Document the run via
write-meta-article-documenting-agent-work– the first documented run defines the pass threshold and seeds the definitive article.
Definition of done (QA checklist)
- Article graded; score and full feedback recorded verbatim in the tracker
- Every flagged issue fixed or explicitly accepted with a written reason
- Fixes verified against transcript and GCT (no invented content)
- Hard Blog Posting Guidelines specs still pass after edits
- Final score attached to the post record; first-run meta-article created
- Linked back to the definitive article and relevant siblings (hub pending)
Example(s)
- Example needed – run the Meta-Article Prompt after first real run. No documented Jennifer grading run exists in the library yet; the first one defines the rubric.
File in the zip: skills/content-factory-process/grade-article-using-jennifer.md. Download the Task Library zip.
Index: Task Library. Article guidelines.
