agentleFS
TheWayWithinTheWayWithin/agent-11/.claude/claude.md

agent-11

CLAUDE.md17 starsChanged 4 months agoMITView on GitHub

This is the AGENT-11 framework repository - a library of AI agents that gets deployed to user projects. When working in this repository, you must understand the critical distinction between two agent directories. Personal preferences: See root /CLAUDE.md for Jamie's communication preferences. When creating library documentation (files in project/field-manual/, project/missions/, etc.), write for general technical users: - Clear, concise technical documentation - Standard professional tone - Assume competent self-directed users - They can follow instructions without hand-holding - No completion…

The file

Show its rules and text

We split this claude.md into the rules it gives an agent

17 rules. About 0 are common, 16 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈16

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    When creating library documentation (files in project/field-manual/, project/missions/, etc.), write for general technical users:

    line 11
  2. Distinctive:

    No completion prompts: - Don't ask "Ready to continue?" in public docs

    line 14
  3. Distinctive:

    Standard formatting: - Use common technical documentation patterns

    line 15
  4. Distinctive:

    ✅ MODIFY THESE: When enhancing agent capabilities for end users

    line 36
  5. Distinctive:

    ✅ MODIFY THESE: When enhancing command capabilities for end users

    line 52
  6. Distinctive:

    Before modifying ANY agent or command file, ask yourself:

    line 64
  7. Distinctive:

    Quick check: Always verify with the deployment script

    line 74
  8. Distinctive:

    If found → it's deployed (library file)

    line 76
  9. Distinctive:

    If not found → it's internal (working squad file)

    line 77
  10. Distinctive:

    Not comprehensively updated (not priority)

    line 168
  11. Distinctive:

    If you realize you've been working on the wrong agent directory:

    line 173
  12. Distinctive:

    Document: Update progress.md with error recovery

    line 179
  13. Distinctive:

    Default to library agents unless explicitly internal work

    line 199
  14. Distinctive:

    Verify target directory before starting significant work

    line 200
  15. Distinctive:

    Document which directory changes apply to

    line 201
  16. Distinctive:

    Update both project-plan.md and progress.md with correct paths

    line 202
  17. Distinctive:

    Critical: Always verify which agents you're modifying before starting work

    line 214

≈ “Common” is an estimate: at least 1 in 5 of the 28,029 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.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
claude.md

CLAUDE.md - AGENT-11 Development Project

PROJECT OVERVIEW

This is the AGENT-11 framework repository - a library of AI agents that gets deployed to user projects. When working in this repository, you must understand the critical distinction between two agent directories.

Personal preferences: See root /CLAUDE.md for Jamie's communication preferences.

Library Documentation Standards

When creating library documentation (files in project/field-manual/, project/missions/, etc.), write for general technical users:

  • Clear, concise technical documentation - Standard professional tone
  • Assume competent self-directed users - They can follow instructions without hand-holding
  • No completion prompts - Don't ask "Ready to continue?" in public docs
  • Standard formatting - Use common technical documentation patterns

⚠️ CRITICAL ARCHITECTURE DISTINCTION

Working Squad (.claude/agents/)

Purpose: Internal development squad for AGENT-11 project work Location: /Users/jamiewatters/DevProjects/agent-11/.claude/agents/ Agent Count: 12 agents Use Case: Building, testing, and maintaining the AGENT-11 framework itself Agents Include: coordinator, developer, strategist, tester, architect, designer, operator, analyst, marketer, support, documenter, agent-optimizer Archived: content-creator (overlaps with marketer), design-review (overlaps with designer)

⚠️ DO NOT MODIFY UNLESS: You are improving the AGENT-11 development process itself

Library Agents (project/agents/specialists/)

Purpose: Deployable agent library for end users Location: /Users/jamiewatters/DevProjects/agent-11/project/agents/specialists/ Agent Count: 11 agents Use Case: Agents that get deployed to user projects via install.sh Agents Include: coordinator, strategist, architect, developer, designer, tester, documenter, operator, analyst, marketer, support

✅ MODIFY THESE: When enhancing agent capabilities for end users

Working Commands (.claude/commands/)

Purpose: Internal slash commands for AGENT-11 development work Location: /Users/jamiewatters/DevProjects/agent-11/.claude/commands/ Use Case: Commands used while building the AGENT-11 framework itself Commands Include: Internal development utilities (NOT deployed)

⚠️ DO NOT MODIFY UNLESS: You are improving the AGENT-11 development process itself

Library Commands (project/commands/)

Purpose: Deployable slash commands for end users Location: /Users/jamiewatters/DevProjects/agent-11/project/commands/ Use Case: Commands that get deployed to user projects via install.sh Commands Include: coord, meeting, design-review, recon, report, pmd, dailyreport, planarchive

✅ MODIFY THESE: When enhancing command capabilities for end users

DEFAULT WORK TARGET

When in doubt, work on:

  • project/agents/specialists/ (Library Agents)
  • project/commands/ (Library Commands)

Reason: Most work involves improving what users get when they deploy AGENT-11, not our internal development tools.

VERIFICATION PROTOCOL

Before modifying ANY agent or command file, ask yourself:

  1. Am I improving the AGENT-11 library for end users?
    • YES → Work on project/agents/specialists/ OR project/commands/
    • NO → Continue to question 2
  2. Am I improving the AGENT-11 development process?
    • YES → Work on .claude/agents/ OR .claude/commands/
    • NO → Ask user for clarification
  3. Quick check: Always verify with the deployment script
    • Run: grep -n "your-file" project/deployment/scripts/install.sh
    • If found → it's deployed (library file)
    • If not found → it's internal (working squad file)

COMMON SCENARIOS

Scenario: "Improve agent prompting"

Target: project/agents/specialists/ (users benefit)

Scenario: "Add new agent capability"

Target: project/agents/specialists/ (users benefit)

Scenario: "Fix bug in /planarchive command"

Target: project/commands/planarchive.md (users benefit)

Scenario: "Enhance /coord command features"

Target: project/commands/coord.md (users benefit)

Scenario: "Optimize agent performance"

Target: project/agents/specialists/ (users benefit)

Scenario: "Fix bug in coordinator delegation"

Target: project/agents/specialists/coordinator.md (users benefit)

Scenario: "Improve our internal code review process"

Target: .claude/agents/ (development process)

Scenario: "Add internal dev utility command"

Target: .claude/commands/ (internal tool, not deployed)

Scenario: "Add agent-optimizer capabilities"

Target: .claude/agents/agent-optimizer.md (internal tool, not deployed)

FILE STRUCTURE REFERENCE

agent-11/
├── CLAUDE.md             ← Jamie's personal preferences (NOT deployed)
├── library/
│   └── CLAUDE.md         ← DEPLOYABLE content → user's .claude/CLAUDE.md
├── .claude/
│   ├── agents/           ← WORKING SQUAD (12 agents, internal use ONLY)
│   ├── commands/         ← Internal slash commands (NOT deployed)
│   └── claude.md         ← This file (project guardrails)
│
└── project/
    ├── agents/
    │   └── specialists/  ← LIBRARY AGENTS (11 agents, DEPLOYED TO USERS)
    ├── commands/         ← LIBRARY COMMANDS (DEPLOYED TO USERS)
    ├── missions/         ← Mission templates (deployed)
    ├── field-manual/     ← User documentation (deployed)
    └── deployment/
        └── scripts/
            └── install.sh ← Deploys library to user projects

DEPLOYMENT SYSTEM

Two-File Architecture (user's project after deployment):

user-project/
├── CLAUDE.md              ← User's personal preferences (NEVER touched)
└── .claude/
    └── CLAUDE.md          ← AGENT-11 library instructions (from library/CLAUDE.md)

What Gets Deployed (see install.sh):

  • library/CLAUDE.md → user's .claude/CLAUDE.md (AGENT-11 instructions)
  • project/agents/specialists/*.md → user's .claude/agents/
  • project/commands/*.md → user's .claude/commands/
  • project/missions/*.md → user's missions/
  • templates/*.md → user's templates/
  • project/field-manual/*.md → user's field-manual/

What Stays Internal (NOT deployed):

  • /CLAUDE.md (root - Jamie's personal preferences)
  • .claude/agents/*.md (working squad)
  • .claude/commands/*.md (internal commands)
  • .claude/claude.md (this file - dev instructions)
  • Development documentation (project-plan.md, progress.md)

PHASE 1 & 2 MODERNIZATION CONTEXT

What Was Modernized (October 2025):

  • ✅ project/agents/specialists/ - 11 library agents fully modernized
  • ✅ Memory tool integration
  • ✅ Extended thinking modes
  • ✅ Tool permissions
  • ✅ Self-verification protocols
  • ✅ Context editing guidance

Working Squad Status:

  • .claude/agents/ received SOME modernization updates
  • Not comprehensively updated (not priority)
  • Focus was on library agents for user benefit

ERROR RECOVERY PROTOCOL

If you realize you've been working on the wrong agent directory:

  1. STOP immediately
  2. Assess: Is the work relevant to library agents?
  3. If YES: Apply changes to project/agents/specialists/ agents
  4. If NO: Revert changes and restart with correct directory
  5. Document: Update progress.md with error recovery

QUICK REFERENCE

TaskTarget Directory
Improve agent capabilitiesproject/agents/specialists/
Add new agent featuresproject/agents/specialists/
Fix agent bugsproject/agents/specialists/
Update agent protocolsproject/agents/specialists/
Optimize agent performanceproject/agents/specialists/
Fix command bugsproject/commands/
Add command featuresproject/commands/
Improve command UXproject/commands/
Improve internal dev process.claude/agents/
Add internal-only tools.claude/agents/ OR .claude/commands/

COORDINATION RULES

When coordinating missions:

  • Default to library agents unless explicitly internal work
  • Verify target directory before starting significant work
  • Document which directory changes apply to
  • Update both project-plan.md and progress.md with correct paths

REMINDER

99% of work should target: project/agents/specialists/

1% of work targets: .claude/agents/ (rare, internal improvements only)


Last Updated: 2025-10-18 Purpose: Prevent confusion between working squad and library agents Critical: Always verify which agents you're modifying before starting work

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.