seo-engine-policies
Skill1 starsChanged 31 days agoMITView on GitHub
States what each search and AI answer engine prohibits, how it enforces, notifies and allows recovery, and which crawler and AI opt-out controls it honours: Google, Bing/Copilot, DuckDuckGo, Ecosia, Yahoo, Qwant, Yandex, Baidu, Naver, Seznam, Brave, OpenAI, Perplexity, Anthropic, Apple. Each rule is cited and dated. Use for "manual action", "is this allowed on Bing", "how do we appeal". Not for whether a tactic works (seo-evidence).
In this repository7 other files it gives its agents.
- seo-ai-searchSkill
- seo-authoritySkill
- seo-contentSkill
- seo-evidenceSkill
- seo-optimizerSkill
- seo-strategySkill
- seo-technicalSkill
The file
We split this skill into the rules it gives an agent
2 rules. About 0 are common, 1 is distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 14
Never forecast a published number onto a client site.
- Very specific: line 15
No rollout without a pre-declared revert rule (applicability, holdout, what stays fixed, metric + window, trigger).
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
seo-engine-policies
What each engine forbids, how it enforces and allows recovery, and which crawler and AI controls it honours, from its dated documents.
<!-- grammar-excerpt -->Label every tactic with one of the 9 marks: [documented: <URL>, <date>] · [documented-secondary: …] · [unverified: …] · [measured: MARK, n=S, c=C] · [practitioner: anonymised sources · …] · [hypothesis] · [vendor] · [memory] · [unvalidated house heuristic].
Numbers keep their denominator; marks are counts, not verdicts; bundles support direction, not size.
Never forecast a published number onto a client site; only the client's own GSC demand and funnel forecast.
No rollout without a pre-declared revert rule (applicability, holdout, what stays fixed, metric + window, trigger).
Repetition never proves; full grammar: skill seo-optimizer § Evidence marks.
<!-- /grammar-excerpt -->Load rule
Open one reference per load-when row the question touches (a "by H2" row: only the part holding that H2); never a file "to be thorough". A pointer into another seo- skill's file is a citation: quote it, or load that skill's SKILL.md and open the file through its load-when row. Engine pages, ledger and practitioner parts and the crawler registry are lookup-only.
Load when
| ID | Question shape | Open |
| EP1 | one named engine, or that engine's own rules | one of: google, google-ai, bing, microsoft-copilot, bing-syndication, yandex, baidu, naver, seznam, brave, openai, perplexity, anthropic, other-ai-crawlers; a dated or regional change ("the EEA change"): google §6, not EP3 |
| EP2 | the answer cites ≥2 engines, or "which engines" / "all engines" | engine-matrix.md, then EP1 per engine the answer quotes (≤3); "which engines matter in a market" is the matrix plus skill seo-strategy › strategy-playbooks, no engine page |
| EP3 | a dated Google update, "was it an update", timeline | by H2: policy and documentation changes or core/spam rollouts, data windows, unverified |
| EP4 | asked what a violation class means | policy-taxonomy.md |
| EP5 | a spam play: run request, or seen in a rival, proposal, tool, offer | by H2: 1, 2, 3, 4, 5, 6, 7, 8, 9 |
Engine index
| Page | Covers | Market relevance |
| Search, Discover, News | global; baseline | |
| google-ai | AI Overviews, AI Mode, Gemini, Google-Extended | global |
| bing | Bing Search | feeds Copilot, DDG, Yahoo, Ecosia (part) |
| microsoft-copilot | Copilot, Bing AI answers | global |
| bing-syndication | DuckDuckGo, Ecosia, Yahoo, Qwant | FR/DE (Qwant, Ecosia) |
| yandex | Yandex | RU, CIS, TR |
| baidu | Baidu | CN |
| naver | Naver | KR |
| seznam | Seznam | CZ |
| brave | Brave Search | global, small |
| openai | ChatGPT search, GPTBot | global AI |
| perplexity | Perplexity | global AI |
| anthropic | Claude | global AI |
| other-ai-crawlers | Apple, Common Crawl, Meta, ByteDance, Amazon | crawlers |
All 14 pages: last_verified 2026-10-11.
How to read a page
13 sections: §1–§4 users, documents, crawlers, tools; §5 prohibited (classes 1–16); §6 enforcement; §7 recovery; §8 AI controls (17–18); §9 delta from Google; §10 practitioner; §11 unverified. A §5/§8 row: engine's term, ≤25-word quote, dated label; "not addressed" is an answer.
Not mine
| Question | Skill |
| does a tactic work, studies, practitioners | seo-evidence |
| fixing robots.txt, canonicals, schema, migrations | seo-technical |
| getting cited by AI engines, edge-gate testing | seo-ai-search |
| link building, E-E-A-T, digital PR | seo-authority |
Rules
- One engine's policy never fills another engine's cell. "Bing probably also…" is
[hypothesis]; "same as Google" needs a cite. - Say the page's
last_verifieddate when quoting; re-fetch if older than 180 days[unvalidated house heuristic]. - An INCOMPLETE page says so, lists what is unverified, names the official channel and gives the safe default; it never invents rules.
- Practitioner penalty reports stay in a page's §10,
[practitioner]. - A forbidden tactic is not graded or built: policy gate before evidence before implementation. A spam play (EP5) is never run, tuned or priced: a run request gets a one-sentence refusal, then the play in entry order, mechanism only; rival, victim, audit: explain fully.
- Default engine set: a rule every engine has (link spam, cloaking, doorways, scaled content) is answered for Google and Bing at least (EP2) when the question names one engine or none; say where others are silent. Site reputation abuse: google §6 only.
- Vanished from an engine: technical causes first (skill
seo-technical› crawl-and-index), then engine page §5–§7; a traffic drop starts at skillseo-technical› incident-triage. - Route on: link penalty: skill
seo-authority› authority-links-and-pr; partner section on your domain: google §6, skillseo-authority› authority-trust-and-distribution, skillseo-content› programmatic-seo-2; lighter page for bots: skillseo-technical› rendering-and-sitemaps-2; mass city pages: skillseo-evidence› works-vs-doesnt, then skillseo-content› programmatic-seo.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.