Fixing four pipeline bugs showed that the blog backlog was a content problem, not a plumbing problem.
Plumbing first
Four bugs were quietly blocking the blog pipeline. Fixing the false positives in the draft funnel cut the blocked count from 217 to 109. I also backfilled the missing lesson blocks across the whole review queue and wrote nineteen drafts for the gap week.
Content was the real blocker
With the plumbing fixed, nothing was publishable for content reasons rather than plumbing. Part of the scoring record also turned out to be untrustworthy.
The 43 posts failing both the opener and closer gates need real editing. Two more were held, one over an ambiguous 0.05 figure and one over a named CVE.
What the numbers say
The plumbing work took the blocked pile from 217 drafts to 109. No bug fix helps the 43 posts that fail both gates, because no tool can write a better opening sentence for them. The only way through them is to go back and rewrite what they say.
AI Skills
Use this lesson with the AI assistant you already use
Four bugs were blocking the blog pipeline. Fixing them cleared about half of the blocked drafts, but 43 posts still failed both the opener and closer gates, because those gates test what a post says.
Paste the prompt, share only the context needed to answer it, and treat the result as a draft for your review. Do not include confidential information or let an AI assistant make changes without your approval.
Optional: for a visual report and saved memory, run /dxdev first.
Don’t have it? Get it at dxdev.com/skills/dxdev. The prompt works without it.
dxdev LESSON ยท paste into your AI coding agent
LESSON: Fix The Plumbing First, Then Check Whether Anything Was Ever Plumbing
SOURCE: dxdev.com/blog/2026-09-24_backlog-was-a-content-problem-not-plumbing
WHAT HAPPENED: A blog pipeline had a large backlog of blocked drafts. Four bugs were fixed, including false positives in the draft funnel, which cut the blocked count from 217 to 109. Missing lesson blocks were backfilled across the review queue and nineteen drafts were written for a gap week. After that, nothing was publishable for content reasons rather than plumbing. 43 posts failed both the opener gate and the closer gate and needed real editing, two more were held for judgment calls, and part of the scoring record turned out to be untrustworthy.
THE RULE: When a backlog stalls, fix the mechanical bugs first, because they hide the real count. Then measure what is left and name the test each remaining item fails. Items that fail a content test cannot be unblocked by more tooling.
CHECK MY CODE, then report PASS or FAIL with file:line for each:
1. Any blocked-item count that has not been checked for false positives in the check that produces it.
2. Any backlog item that fails a content gate and is being retried through tooling instead of being edited.
3. Any score or verdict record that is used to make publish decisions without being spot-checked against the item itself.
THEN PRINT: a table (check, PASS/FAIL, evidence, fix) + a verdict (applies / partially / OUT_OF_SCOPE / no) + the single most important next action.