Cursor rules
The most expressive format here. Many small .mdc files, each with frontmatter declaring when it should load, so a rule about migrations only enters context when a migration is open. Costs the most to maintain and only one editor reads it.
What the corpus measured
2026-08-03Repositories
179
2026-08-03Adopting repositories, not files.Share of corpus
25.8%
2026-08-03Of every repository RuleStack indexes.Median length
272w
2026-08-0390th percentile 1,171 words.Median quality
52
2026-08-0329.3% name a command; 58.0% contain code.Repositories carrying Cursor rules, 2026-08-01 to 2026-08-03
Real Cursor rules files
| Repository | Format | Stack | Covers | Score | Changed |
|---|---|---|---|---|---|
| hiromaily/go-crypto-wallet.cursor/rules/typescript.mdc · 126 | Cursor rules | setupbuildtestlint-format+6 | 100/100 | 2 days ago | |
| TechSquidTV/Hermes.cursor/rules/10-hermes-api.mdc · 45 | Cursor rules | testlint-formatstylearch+5 | 100/100 | 2 days ago | |
| langflow-ai/langflow.cursor/rules/docs_development.mdc · 153k | Cursor rules | setupbuildtestlint-format+7 | 97/100 | 2 days ago | |
| TechSquidTV/Hermes.cursor/rules/20-hermes-api-tests.mdc · 45 | Cursor rules | teststyletesting-strategysecurity+3 | 97/100 | 2 days ago | |
| skillrecordings/egghead-next.cursor/rules/gh-task-plan.mdc · 1.4k | Cursor rules | teststylearchtypes+2 | 96/100 | 2 days ago | |
| skillrecordings/egghead-next.cursor/rules/project-update-rules.mdc · 1.4k | Cursor rules | buildtestlint-formatstyle+7 | 96/100 | 2 days ago | |
| skillrecordings/egghead-next.cursor/rules/project-update-user-rules.mdc · 1.4k | Cursor rules | buildtestlint-formatstyle+7 | 96/100 | 2 days ago | |
| hiromaily/go-crypto-wallet.cursor/rules/proto.mdc · 126 | Cursor rules | buildlint-formatstylearch+3 | 96/100 | 2 days ago | |
| nerds-odd-e/doughnut.cursor/rules/cli.mdc · 49 | Cursor rules | setupbuildteststyle+4 | 96/100 | 2 days ago |
What it asserts
2026-08-02Read from the format’s own documentation and checked by hand. Capabilities are asserted; adoption is measured.
- File
- .cursor/rules/*.mdc
- When it applies
- Four modes set in frontmatter: Always Apply (every chat session), Apply Intelligently (the agent picks it by description), Apply to Specific Files (globs match the edited file), and Apply Manually (@-mentioned). Rules must carry the .mdc extension — a plain .md file in .cursor/rules is ignored.
- Size limits
- None documented
- Tools that read it
- Cursor
Capabilities
- Read by more than one tool
- Frontmatter
- Glob targeting
- Imports other files
- Nested / per-directory
- Multiple files
- User-level scope
Strengths
- Glob-scoped activation — the only widely-used format that keeps unrelated rules out of context
- Description-based activation lets the agent pull a rule in on demand
- Many small files are easier to review in a PR than one growing document
Weaknesses
- Cursor-only
- Frontmatter is a schema, and a malformed .mdc silently stops applying
- Fragmentation: rules spread over many files are easy to duplicate and contradict
Measured — counted from real repositories, nightlyAsserted — read from the vendor docs, dated
Repositories that carry Cursor rules and something else
Counted pairwise across the whole corpus. The percentage is of repositories carrying Cursor rules — the number a “Cursor rules vs …” search is really asking for.
+ AGENTS.md
38.0%
2026-08-0368 repos carry both17.2% the other way
+ CLAUDE.md30.7%
2026-08-0355 repos carry both22.2% the other way
+ Copilot instructions16.2%
2026-08-0329 repos carry both23.0% the other way
+ Cline rules7.8%
2026-08-0314 repos carry both17.7% the other way
+ .cursorrules7.3%
2026-08-0313 repos carry both48.1% the other way
+ Windsurf rules6.7%
2026-08-0312 repos carry both85.7% the other way
+ GEMINI.md6.1%
2026-08-0311 repos carry both33.3% the other way
