jev-seo
AGENTS.md100 starsChanged 9 days agoMITView on GitHub
Where files are headingNew AGENTS.md files have more rules about keeping secrets out of the code, and fewer about code style.
Of the AGENTS.md files we read, the 31,700 first seen in the last six months, against the older ones: the kind of rule that gained most share, and the kind that lost most. About files like this in general, not this one, and an estimate, since a kind is a pattern in the words.
In this repository3 other files it gives its agents.
- site/llms.txtllms.txt
- jev-seoSkill
- jev-auditSkill
The file
We split this AGENTS.md into the rules it gives an agent
3 rules. About 0 are common, 2 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
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
Humanizer & Stop-Slop: All public docs, READMEs, and commit messages must strictly avoid AI vocabulary, passive voice, and boilerplate.
- Very specific: line 16
Commit locally, push on "go": never push master or site without an explicit go.
- Distinctive: line 19
Every marketing page or documentation footer must reference:
≈ “Common” is an estimate: at least 1 in 5 of the 40,246 AGENTS.md 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
Project Rules & Agent Instructions: jev-seo
Overview
jev-seo is a 100% Free and Open Source (FOSS), agent-first SEO & Generative Engine Optimization (GEO) suite powered by TypeSafe AI Jev (System One) and local zero-cost scraping. An open-source, subscription-free alternative to Semrush, Ahrefs, and OpenSEO.
- Registry: 64 stable rules (R01-R64) across 9 areas; 142 tests green; rule-set build
2026.09b. - Free path costs ₹0 (DuckDuckGo + local files). Jev semantic calls are hard-capped (
--jev-budget, default $0.25/run;--no-jevis fully offline). Paid backends (Tavily, DataForSEO, Jina, Firecrawl) are explicit opt-in only: flag + key, never silent. - v0.4.0 unreleased (Search CI:
check+verify,.jev-seo/policy.toml, regional SERP, GSC inspect/winners/losers, Bing sites/queries, SARIF/JUnit, R59-R64, R37 informational, citation struct, indexability table, 4 new MCP tools,jev://resources, evidence snapshots, portable manifest).v0.3.0shipped 2 October 2026 (64 rules, 142 tests).v0.1.2shipped 26 September 2026. Former unreleased0.1.3was merged intov0.3.0.
Core Directives
- FOSS & Zero-Cost Free Path: MIT licensed. DuckDuckGo + local files + SQLite drift are free forever. No paid API is required; DFS/Tavily/Jina/Firecrawl exist only behind opt-in flags.
- TypeSafe Jev System One First: Use deterministic primitives (
Choice,Score,Noul) for intent classification, content gap analysis, and GEO citation scoring. Never use Jev for arithmetic, character counts, or generative prose. Choice/Score act at confidence 0.80; anything under prints[verify]or a withheld number, never a confident fake. - Speculative Fan-out: Ingest state once, batch independent questions into a single request.
- Hardware & Resource Limits: Optimize for low-RAM (Windows 11, 8GB total memory). Fast cold starts, streaming output, zero heavy headless browser dependencies by default. (Raw-vs-rendered DOM diff is scoped as its own future version, not bolted on.)
- Humanizer & Stop-Slop: All public docs, READMEs, and commit messages must strictly avoid AI vocabulary, passive voice, and boilerplate.
- Gates over guesses: 22 rules are Blocking (fail CI via
--fail-on), 36 are Advisory warnings.fact/heuristictruth tags on every finding; heuristic rules never fail a build silently. - Commit locally, push on "go": never push
masterorsitewithout an explicit go.[skip ci]on docs/asset-only commits.
Ecosystem Footer Standard
Every marketing page or documentation footer must reference:
- Ecosystem:
jev-superpowers,jev-curate,tdlib-android,kharcha - Author: Akash Priyadarshi (Patna, Bihar, India)
- Links: Portfolio (https://akashpriyadarshi.vercel.app), GitHub (https://github.com/AkashPriyadarshii)
Live Site Deploy
- Site files live in
master/site/(vendored,linguist-vendored). GitHub Pages is off; Vercel servessite/as canonical. Editsite/onmaster; push on go. - GitHub Pages: https://akashpriyadarshii.github.io/jev-seo/ auto-deploys on push to
site. - Vercel: https://jevseo.vercel.app (user-managed link, independent of Pages).
- Edit flow: edit
site/onmaster, commit, push on go.masterstays pure Rust crate plus vendored site.
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.

