fig — SEO/AEO strategy skill

What it does

Here, agent means the chat or coding assistant you already use (Cursor, Claude, Grok, Codex). Fig is not SEO for talent agents, sports agents, or real-estate agents.

On first-run after Checkout or a funded balance, the fig skill posts the intro, then takes the site/goal with Automatic mode as the default. Automatic mode researches and implements so the User reviews landed work; implementation goes through host subagents, new bots, or equivalent sessions. Strategic mode researches first, iterates until the User agrees the plan, and stops at the approved plan. Both modes use the Selection rubric: repair, refresh, consolidate, and prune beat create when evidence is close. Pull fig-seo-setup for needs-you and the SEO connections ladder; pull fig-content-strategy for what to publish and for census (prune, merge, refresh); pull fig-content-craft for polish, visuals, and original-research pages; pull fig-programmatic-seo for gated template sets; pull fig-free-tool when a calculator or generator beats a guide; pull fig-offpage for off-page briefs; pull fig-distribute after a URL ships; pull fig-outcomes to register and score shipped actions.

Runs read and write lite operator memory in the User's project (fig/operator-memory.json): site, market, vitals, last actions, open priorities, observation windows, and baselines — observed values only. A quiet measurement is a valid Automatic mode outcome when nothing beats the current bar.

Specialists

• fig-seo-setup — needs-you queue and the SEO connections ladder (GSC, analytics, CMS); not Fig Checkout.
• fig-site-audit — technical, crawlability, indexability, on-page, entity, and Core Web Vitals / performance baseline; reads and writes operator memory on audit handoff.
• fig-search-research — demand, intent, SERPs, competitors, and keyword opportunities.
• fig-answer-research — AI visibility, citability, structured data, and answer-engine access.
• fig-content-strategy — prioritization, information architecture, content census (prune, merge, refresh), and content briefs.
• fig-content-craft — polish, AI-writing tells, visuals/tables, and proprietary-data page patterns.
• fig-programmatic-seo — gated compare, alternatives, and for-persona template sets.
• fig-free-tool — gated calculator, checker, and generator briefs when a tool beats a guide.
• fig-content-validation — evidence-backed checks after a change ships.
• fig-outcomes — register a shipped action and score it after the observation window.
• fig-offpage — directories, digital PR angles, and unlinked-mention briefs. Fig does not send outreach.
• fig-distribute — post-publish channel checklist and timed rechecks after a URL ships.

• Canonical SKILL.md
• Operator memory shape

Usage and evidence

Fig uses prepaid credits. Checkout starts at $10; $20 is the recommended pack. Your agent spends the balance on live lookups and asks you to approve Checkout again when it needs more.

Use a prioritized research set and make metrics, mentions, citations, and observed movement explicit in the result. When Search Console is connected, name GSC baselines from returned rows; when it is not, mark Search Console measurement unavailable and do not invent GSC numbers.

Related

• fig-seo-setup
• fig-site-audit
• fig-search-research
• fig-answer-research
• fig-content-strategy
• fig-content-craft
• fig-programmatic-seo
• fig-free-tool
• fig-content-validation
• fig-outcomes
• fig-offpage
• fig-distribute
• Fig MCP
• All skills
• Agent guides
• Agent SEO
• AEO for agents
• Fig tools
• Compare Fig