cursor-kit
AGENTS.md0 starsChanged 55 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 31,774 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 repository27 other files it gives its agents.
- claude/CLAUDE.mdCLAUDE.md
- ai-memory-curateSkill
- asd-ste100Skill
- blindspot-passSkill
- code-reviewSkill
- diagnosing-bugsSkill
- digest-filesSkill
- domain-modelingSkill
- emodeSkill
- find-skillsSkill
- grillingSkill
- handoffSkill
- new-botSkill
- prior-art-checkSkill
- prototypeSkill
- prSkill
- researchSkill
- resolving-merge-conflictsSkill
- retroSkill
- setup-ai-memorySkill
- setup-matt-pocock-skillsSkill
- setup-projectSkill
- setup-pucprSkill
- simplifySkill
- tddSkill
- verification-planningSkill
- wayfinderSkill
The file
We split this AGENTS.md into the rules it gives an agent
5 rules. About 0 are common, 2 are distinctive, and 3 are 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.
3
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 6
Prefer minimal tokens: caveman_stats on demand only
- Very specific: line 7
Prefer /setup-project for new or existing repos on this stack
- Very specific: line 9
Prefer /poteto-mode for non-trivial kit packaging/architecture work
- Distinctive: line 16
Prefer impartial 0–100 rankings when comparing this stack to peer toolkits.
- Very specific: line 40
Never store secrets (profile/habits/never-store-secrets.md) <!-- ai-memory:profile:end -->
≈ “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
AGENTS.md — cursor-kit
Learned User Preferences
- Cursor-only dev environment (macOS primary; Windows scripts still shipped)
- Prefer minimal tokens: caveman_stats on demand only
- Prefer
/setup-projectfor new or existing repos on this stack /setup-projectlocked decisions: A2 (per-repo + global repair if broken), B3 (existing or new — ask), interview depth 1 minimal vs 2 completo — ask; ship as plugin inside cursor-kit- Prefer
/poteto-modefor non-trivial kit packaging/architecture work - Human-facing kit docs may be PT; pasteable prompts stay English; avoid absolute machine-specific paths in shared docs
- New kit skills should be prompt-driven orchestrators (compose existing skills/templates), not monoliths; draft SKILL.md for approval before running setup against real repos
- Steal onboarding/playbook/quality-gates/skill form from peer toolkits when useful; create native skills (do not vendor third-party SKILL.md verbatim); wire into
setup-project/setup-pucpr; keep stack advantages (token-engine, hooks +.cursor/state,/setup-project, ponytail/caveman/cbm, continual-learning, pstack/team-kit) - Quality gates are first-class: measure-only vs autofix stay separate;
/setup-projectdepth 2 may offer measure-only gates and optional ai-memory companion - Cursor-primary workflow: cursor-kit plugin is the sole skill SoT on desktop (
~/.cursor/plugins/local/cursor-kit/skills/). Keep~/.cursor/skillsempty. Do not keep kit or promoted skill copies under~/.cursor/skills,~/.agents/skills, or Agent Store. Codex is occasional only; if used, prefer the plugin path over local skill trees - Long-horizon memory: optional ai-memory companion (
/setup-ai-memory); kit AGENTS/PROJECT/checkpoint stay authoritative for prefs/decisions/live task - Prefer impartial 0–100 rankings when comparing this stack to peer toolkits; do not self-favor
Learned Workspace Facts
- Kit lives at
~/Projects/cursor-kit(git checkout, no symlink; GitHubBossmann007/cursor-kit) - Working branch for Enzo's personal kit config:
enzo(default shipping branch remainsmasteruntil merged) - Pairs with token-engine at
~/Projects/token-engine(GitHubBossmann007/token-engine) - Hooks + rules source of truth:
cursor-kit/hooks/+hooks.json.template→sync-hooks.sh/.ps1(merge-safe; preserves companion hooks);cursor-kit/rules/(+ optionalsync-rules.sh); pstack seed attemplates/pstack/pstack-models.mdc.example - Plugin entry:
.cursor-plugin/plugin.jsonwithskills,rules,hooks(setup orchestrators includingsetup-matt-pocock-skills, quality skills, and promoted engineering skills underskills/); desktop sync via rsync →~/.cursor/plugins/local/cursor-kit/; keep~/.cursor/skillsand Agent Store empty of skill copies - Companion plugins (UI install): continual-learning, cursor-team-kit, pstack as usage companion (skills stay in pstack plugin — not vendored into cursor-kit; then
/setup-pstack); inventory indocs/EXTERNAL-COMPONENTS.md - Optional long-horizon companion: ai-memory via
scripts/install-ai-memory.sh//setup-ai-memory(native macOS + LaunchAgent);sync-hooksmerge-safe so companion hooks survive kit sync - Stranger onboarding docs:
docs/00-overview.md,docs/02-playbook-onboarding.md,docs/prompts/,templates/quality-gates/ - Cursor Projects (Cloud) VMs should use the plugin or
scripts/install-cloud-skills.sh/templates/cloud/environment.jsononly when the plugin is absent (script skips user-skills copy when local plugin is already present) - Course folders default under
~/PUCPR; academic setup is composed viasetup-pucpr(does not replace/setup-project) - Matt
/teach(mission/lessons workspace) and pstack/teach(code explain via how/why) are different skills — do not conflate in wiring - Removed unused kit env skills (
context-engine,dev-workflow,project-brain,update-checkpoint); their intent lives in rules/hooks//setup-projectinstead
<!-- ai-memory:profile:start -->Usual choices (ai-memory profile)
Written by ai-memory profile apply from the user's cross-project profile. Change the profile (ai-memory profile show / forget) and re-run the command instead of editing this block. Everything else in this file takes precedence.
- [javascript] JavaScript is one of the stacks this user usually works in. (
profile/stack/javascript.md) ## Agent skillsblock in existingAGENTS.mdorCLAUDE.md(never create the other when one exists; (profile/habits/agent-skills-block-in-existing-agents-md-or-claude-md-never.md)- Never store secrets (
profile/habits/never-store-secrets.md)
<!-- ai-memory:profile:end -->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.

