Tag
Start here
In our 20 prompt test ChatGPT searched on 8 prompts and answered the other 12 from memory with no sources at all. Here is a six step check that separates a blank result from a real zero, and the agent job that keeps it running.
Tag
Build a Codex skill that runs the operating loop of one X account: pick today's slot from a logged content mix, draft in the right shape, reject weak hooks with a deterministic gate, and review performance from an X API pull or a manual export.
A Search Console export is not a ranking report. This is the weekly workflow we use to turn query data into a report that says what moved, what probably caused it, and what to check next — built once with Codex and rerun in minutes.
OpenAI announced GPT-6 Astra on September 3, 2026, with API access for enterprise Trusted Access customers the same day. The model's pixel-level computer use and site-and-document generation turn agentic SEO from theory into a browser-native task list.
OpenAI began rolling out GPT-6 Astra on September 1, 2026 with Codex among the confirmed surfaces. See what is confirmed, what remains unannounced, and how to set up an automated SEO workflow.
A ready-to-run Codex skill that composes every installed codex-seo-* skill into one audited site: probes the site, inventories the skills, runs each single-URL scanner, builds the weighted 7-category board, writes the report artifacts - and withholds the health score when any category has no measured source.
A ready-to-run Codex skill that recons a live site first (title, H1s, JSON-LD, robots, sitemap inventory), then builds a measured SEO strategy: six industry templates, four phased roadmaps with exit gates, and a KPI table where every baseline row has a real source.
A ready-to-run Codex skill that fetches competitor pricing pages, extracts verifiable facts (prices, plans, ratings, schema, trust signals), and builds a comparison matrix where every row carries its source URL and an as-of date.
A Codex skill and zero-dependency scanner for pages at scale: it groups your sitemap into template families, measures how unique each family really is against the 40% gate and 30% hard stop, and reports five scored categories with the evidence lines left in.
Get the codex-seo-sxo skill, read a Google SERP backwards for the page type it rewards, and watch it give one page two verdicts in one day: ALIGNED on its commercial keyword, MISMATCH on its informational one.
A zero-dependency Codex skill that scores product pages on six dimensions and validates Product JSON-LD against Google Merchant requirements, demonstrated on a bundled demo store going from 17/100 to 98/100.
A one-page hreflang audit that checks self-references, return tags, language codes, and content parity across language versions, with a working Python checker you can install into Codex in two minutes.
Cluster keywords by shared Google results, not text similarity. This article installs a Codex skill that scores SERP overlap and designs hub-and-spoke content architecture.
