claude-code / segment_3_agents
CLAUDE.md223 starsChanged 4 months agoMITView on GitHub
This file is the lesson. Claude reads it on every turn when working under segment_3_agents/. Edit the markdown, change the autonomous behavior. No restart required.
- Reads credentials
In this repository14 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- data/CLAUDE.mdCLAUDE.md
- mcp-teaching-kit-main/CLAUDE.mdCLAUDE.md
- segment_1_quickstart/CLAUDE.mdCLAUDE.md
- segment_2_context/CLAUDE.mdCLAUDE.md
- segment_4_hero/CLAUDE.mdCLAUDE.md
- .github/copilot-instructions.mdCopilot instructions
- azure-ai-deploySkill
- azure-bicep-skillSkill
- claude-md-auditSkill
- genai-prompt-evalSkill
- kubernetes-patternsSkill
- mcp-scaffoldSkill
- review-changesSkill
The file
We split this CLAUDE.md into the rules it gives an agent
8 rules. About 0 are common, 2 are distinctive, and 6 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.
6
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 7
Read any file in segment_3_agents/ or its siblings.
- Very specific: line 9
Create new example files under segment_3_agents/examples/ if the user requests one.
- Very specific: line 10
Run npx tsx segment_3_agents/<file>.ts to validate a demo compiles.
- Very specific: line 17
Run destructive git commands: reset --hard, push --force, branch -D, clean -fd.
- Very specific: line 18
Delete files. If something looks wrong, rename it with a .bak suffix and report it.
- Very specific: line 24
After any edit, run npx tsx <edited-file> to confirm it still parses.
- Distinctive: line 28
Stop immediately. Do not attempt to "fix" by making more changes.
- Distinctive: line 31
Wait for human approval before the next action.
≈ “Common” is an estimate: at least 1 in 5 of the 26,980 CLAUDE.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
Agent Boundaries - Segment 3 Working Directory
This file is the lesson. Claude reads it on every turn when working under
segment_3_agents/. Edit the markdown, change the autonomous behavior. No restart required.
What you CAN do
- Read any file in
segment_3_agents/or its siblings. - Edit the demo TypeScript files (
02_agent_loop.ts,03_agent_boundaries.ts). - Create new example files under
segment_3_agents/examples/if the user requests one. - Run
npx tsx segment_3_agents/<file>.tsto validate a demo compiles.
What you CANNOT do
- Modify
tsconfig.json,package.json, orpackage-lock.json. Other segments depend on the toolchain staying still. - Edit
.env,.env.example, or any file containing secrets. - Touch files outside
segment_3_agents/without an explicit instruction naming the target file. - Run destructive git commands:
reset --hard,push --force,branch -D,clean -fd. - Delete files. If something looks wrong, rename it with a
.baksuffix and report it.
Before making changes
- Confirm you are on a feature branch, not
main. If onmain, stop and report. - State the change you are about to make in one sentence, then make it.
- After any edit, run
npx tsx <edited-file>to confirm it still parses.
If something goes wrong
- Stop immediately. Do not attempt to "fix" by making more changes.
- Report exactly what failed, including the command and the error.
- Suggest the smallest reversible step (usually
git diffthengit checkout -- <file>). - Wait for human approval before the next action.
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.

