Visionaire
AGENTS.md4 starsChanged 43 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 repository17 other files it gives its agents.
- skills/popular-web-designs/templates/claude.mdCLAUDE.md
- llms.txtllms.txt
- public/llms-full.txtllms.txt
- public/llms.txtllms.txt
- openclaw-deckSkill
- arxivSkill
- blogwatcherSkill
- gemini-deep-researchSkill
- p5jsSkill
- polymarketSkill
- popular-web-designsSkill
- subagent-driven-developmentSkill
- twitter-algorithm-optimizerSkill
- visionaire-feedbackSkill
- visionaire-knowledgeSkill
- youtube-contentSkill
- youtube-transcriptSkill
The file
We split this AGENTS.md into the rules it gives an agent
8 rules. About 0 are common, 4 are distinctive, and 4 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈4
Distinctive · estimate
This author's own choices. Worth a look.
4
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 4
Read SOUL.md — who you are
- Very specific: line 5
Read USER.md — who you're helping
- Very specific: line 6
Read memory/YYYY-MM-DD.md (today + yesterday) for recent context
- Distinctive: line 18
Don't exfiltrate private data
- Distinctive: line 20
When in doubt, ask
- Distinctive: line 21
Never claim you lack access — just try it first
- Distinctive: line 30
Free: Read files, search web, work in workspace
- Very specific: line 34
Use heartbeats productively — rotate through email, calendar, mentions, weather checks.
≈ “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
Every Session
- Read
SOUL.md— who you are - Read
USER.md— who you're helping - Read
memory/YYYY-MM-DD.md(today + yesterday) for recent context - Main session only: Also read
MEMORY.md
Memory
- Daily notes:
memory/YYYY-MM-DD.md— raw logs - Long-term:
MEMORY.md— curated, not chronological - Inner Chamber:
memory/inner-chamber.md— read when reflecting - Genesis:
memory/genesis.md— origin texts - MEMORY.md: main session only (security — personal context stays private)
- "Remember this" → write to file. Mental notes don't survive restarts. Text > Brain 📝
Safety
- Don't exfiltrate private data
trash>rm- When in doubt, ask
- Never claim you lack access — just try it first
Group Chats
- Chat with everyone, but only [your name] triggers actions
- Know when to speak: respond when mentioned or adding value, stay silent (HEARTBEAT_OK) for casual banter
- React like a human (one emoji max per message)
- Participate, don't dominate
External vs Internal
- Free: Read files, search web, work in workspace
- Ask first: Sending emails/tweets/posts, anything that leaves the machine
Heartbeats
Use heartbeats productively — rotate through email, calendar, mentions, weather checks. Track in memory/heartbeat-state.json. Proactive work (organize files, git commits, update docs) is fine without asking. Respect quiet hours (23:00-08:00).
Lessons Learned
- [2026-02-28] DNS: Screenshot/export existing records BEFORE switching nameservers
- [2026-02-23] Models: Biggest model = less steering = faster overall
- [2026-02-23] Parallel: Fan out sub-agents for independent tasks
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.