name: ai-coding-daily-writer description: Draft and maintain daily AI coding trend articles for a docs or blog repository. Use when Codex needs to research the latest AI coding models, agents, IDEs, release notes, pricing changes, workflow shifts, or benchmark updates; create a dated daily brief; or turn current product changes into a publishable article.
AI Coding Daily Writer
Overview
Create publishable daily AI coding articles with a fixed structure, dated sourcing, and clear engineering impact. Use the bundled script to open a dated blog file, then fill it with verified, time-sensitive content.
Workflow
1. Define scope
- Use the requested date if the user provides one. Otherwise use the current local date.
- Keep scope tightly focused on AI coding, not general AI news.
- Prefer one daily brief over multiple fragmented updates.
2. Research before writing
- Browse because the content is time-sensitive.
- Prefer primary sources: official blogs, docs, release notes, GitHub releases.
- Use secondary analysis only to add context, not as the factual base.
- Read
source-playbook.mdbefore drafting if the topic mix is unclear.
3. Create the daily file
From the repo root, run:
./skills/ai-coding-daily-writer/new-daily-brief.sh <topic-slug>
The script creates:
blog/YYYY-MM-DD-<topic-slug>.md
Use short ASCII slugs such as codex-cli-release-shift or claude-code-team-rollout.
4. Draft the article
Use this structure:
TL;DRWhat changed todayWhy it mattersWhat to testWatchlistSources
Keep the piece focused on 3 to 5 material changes.
5. Preserve rigor
- Write exact dates for time-sensitive facts.
- Distinguish facts from inference.
- Translate product announcements into engineering consequences.
- If evidence is thin, say so explicitly.
- If nothing truly important happened, publish a short no-signal brief instead of forcing weak items.
Resources
source-playbook.md: source priority, topic selection rules, and article checklistnew-daily-brief.sh: creates a Docusaurus blog post template for the current date
Output Standard
- The article must be ready to publish in a Docusaurus blog without extra restructuring.
- Keep tone analytical, not promotional.
- Every item should answer one question: what should an engineer or technical leader do with this information?