The agent config corpus
Real instruction files from live repositories, re-counted nightly and scored.
4,490 instruction files · 695 live repositories · last run yesterday · 1,196 repositories visited
| Repository | Format | Stack | Covers | Score | Changed |
|---|---|---|---|---|---|
| n8n-io/n8npackages/@n8n/agents/AGENTS.md · 199k | AGENTS.md | buildteststylearch+3 | 100/100 | 3 days ago | |
| pytorch/pytorch.github/copilot-instructions.md · 102k | Copilot instructions | setupbuildteststyle+5 | 100/100 | 3 days ago | |
| mui/material-uiAGENTS.md · 99k | AGENTS.md | setupbuildtestlint-format+9 | 100/100 | 3 days ago | |
| microsoft/playwrightCLAUDE.md · 94k | CLAUDE.md | buildtestlint-formatstyle+7 | 100/100 | 3 days ago | |
| louislam/uptime-kuma.github/copilot-instructions.md · 90k | Copilot instructions | setupbuildtestlint-format+9 | 100/100 | 3 days ago | |
| vllm-project/vllmAGENTS.md · 88k | AGENTS.md | setuptestlint-formatstyle+5 | 100/100 | 3 days ago | |
| elastic/elasticsearchx-pack/plugin/core/src/main/java/org/elasticsearch/xpack/core/ml/AGENTS.md · 78k | AGENTS.md | buildtestlint-formatstyle+2 | 100/100 | 3 days ago | |
| elastic/elasticsearchx-pack/plugin/inference/AGENTS.md · 78k | AGENTS.md | buildtestlint-formatstyle+3 | 100/100 | 3 days ago | |
| code-yeongyu/oh-my-openagentpackages/web/AGENTS.md · 67k | AGENTS.md | setupbuildtestlint-format+6 | 100/100 | 2 days ago | |
| rails/railsAGENTS.md · 59k | AGENTS.md | teststylearchgit+4 | 100/100 | 3 days ago |
Recently changed
Newest first| Repository | Format | Stack | Covers | Score | Changed |
|---|---|---|---|---|---|
| arturogarrido/claudinhoAGENTS.md · 27 | AGENTS.md | setuptestlint-formatstyle+8 | 84/100 | yesterday | |
| arturogarrido/claudinho.cursor/rules/trust-boundary.mdc · 27 | Cursor rules | do-not | 58/100 | yesterday | |
| arturogarrido/claudinho.cursor/rules/change-discipline.mdc · 27 | Cursor rules | no sections | 30/100 | yesterday | |
| diegosouzapw/OmniRouteCLAUDE.md · 38k | CLAUDE.md | setupbuildtestlint-format+9 | 84/100 | yesterday | |
| bgreenwell/dotagentsAGENTS.md · 80 | AGENTS.md | no sections | 25/100 | yesterday | |
| localsend/localsendAGENTS.md · 87k | AGENTS.md | buildtestlint-formatstyle+2 | 84/100 | yesterday | |
| ossrs/srs.claude/CLAUDE.md · 29k | CLAUDE.md | monorepoagent-behaviour | 16/100 | yesterday | |
| ossrs/srs.agents/AGENTS.md · 29k | AGENTS.md | monorepoagent-behaviour | 16/100 | yesterday |
Format adoption
Full comparisonShare of indexed repositories carrying each format. One Cursor setup with thirty .mdc files is one adopting repository, not thirty.
57.0%
2026-08-0335.7%
2026-08-0325.8%
2026-08-0318.1%
2026-08-034.8%
2026-08-032.0%
2026-08-0311.4%
2026-08-033.9%
2026-08-03The corpus, in numbers
2026-08-03Instruction files
4,490
2026-08-03Indexed and publishable. Stubs and template folders never enter.Repositories
695
2026-08-03Adoption share counts repositories, never files.Most adopted
57.0%
2026-08-03of repositories carry AGENTS.mdFiles changed last run
20
2026-08-03Freshness is the moat, so it is also the health check.Nobody else publishes the counts
Every comparison page on the web is a capability table someone wrote from the vendor docs in an afternoon. None of them can tell you how many real repositories ship each file, how long the median one runs, or how often two of them turn up in the same project. Those answers need a corpus, and RuleStack counts one every night.
Discovery runs against live repositories nightly. Every file is classified by stack from declared dependencies, config files, repository language and topics — never from a passing mention in the prose — and every assigned tag keeps the evidence that earned it. Quality scores the file, not the repository, so a one-line AGENTS.md in a 100k-star project does not outrank a genuinely good one in a small library.
Capabilities are asserted; adoption is measured.
