agentleFS
catlog22catlog22/Claude-Code-Workflow/.codex/AGENTS.md

Claude-Code-Workflow

AGENTS.md2.1k starsChanged 7 months agoMITArchived repositoryView on GitHub

Incremental Progress: - Small, testable changes - Commit working code frequently - Build on previous work (subtasks) Evidence-Based: - Study 3+ similar patterns before implementing - Match project style exactly - Verify with existing code Pragmatic: - Boring solutions over clever code - Simple over complex - Adapt to project reality Context Continuity (Multi-Task): - Leverage resume for consistency - Maintain established patterns - Test integration between subtasks Git Operations (Parallel Task Safety): - Only stage/commit files directly produced by…

1 other project uses a version of this.

Adopted by

Most recently taken up by

The file

Show its rules and text

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

22 rules. About 0 are common, 15 are distinctive, and 7 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈15

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 7

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Follow project's existing patterns

    line 7
  2. Distinctive:

    Test all public functions

    line 14
  3. Distinctive:

    Test edge cases and error conditions

    line 15
  4. Distinctive:

    Don't expose sensitive info

    line 23
  5. Distinctive:

    Commit working code frequently

    line 29
  6. Distinctive:

    Verify with existing code

    line 35
  7. Distinctive:

    Test integration between subtasks

    line 45
  8. Distinctive:

    Only stage/commit files directly produced by current task

    line 48
  9. Distinctive:

    Never touch unrelated changes or other task outputs

    line 49
  10. Very specific:

    Use git add <specific-files> instead of git add .

    line 50
  11. Distinctive:

    Verify staged files before commit to avoid cross-task conflicts

    line 51
  12. Distinctive:

    If your task conflicts with existing uncommitted changes, STOP and report the conflict instead of overwriting

    line 54
  13. Distinctive:

    Treat all pre-existing uncommitted changes as intentional work-in-progress by other tools

    line 55
  14. Very specific:

    Direct Binary Calls: Always call binaries directly in functions.shell, set workdir, avoid shell wrappers (bash -lc, cmd /c, etc.)

    line 60
  15. Very specific:

    Use apply_patch tool for all routine text edits

    line 63
  16. Distinctive:

    Avoid Python editing scripts unless both fail

    line 65
  17. Distinctive:

    Use for: finding implementations, understanding architecture, locating patterns

    line 89
  18. Very specific:

    Use smart_search(query="...") for keyword/regex search

    line 93
  19. Very specific:

    Use smart_search(action="find_files", pattern="*.ts") for file discovery

    line 94
  20. Very specific:

    Read multiple files in parallel: read_file(path="file1.ts"), read_file(path="file2.ts")

    line 98
  21. Very specific:

    NEVER: use shell commands (cat, find, grep) when MCP tools are available.

    line 106
  22. Distinctive:

    Before starting work, scan recent sessions (7 days) to avoid conflicts and reuse prior work:

    line 119

≈ “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.

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

Codex Code Guidelines

Code Quality Standards

Code Quality

  • Follow project's existing patterns
  • Match import style and naming conventions
  • Single responsibility per function/class
  • DRY (Don't Repeat Yourself)
  • YAGNI (You Aren't Gonna Need It)

Testing

  • Test all public functions
  • Test edge cases and error conditions
  • Mock external dependencies
  • Target 80%+ coverage

Error Handling

  • Proper try-catch blocks
  • Clear error messages
  • Graceful degradation
  • Don't expose sensitive info

Core Principles

Incremental Progress:

  • Small, testable changes
  • Commit working code frequently
  • Build on previous work (subtasks)

Evidence-Based:

  • Study 3+ similar patterns before implementing
  • Match project style exactly
  • Verify with existing code

Pragmatic:

  • Boring solutions over clever code
  • Simple over complex
  • Adapt to project reality

Context Continuity (Multi-Task):

  • Leverage resume for consistency
  • Maintain established patterns
  • Test integration between subtasks

Git Operations (Parallel Task Safety):

  • Only stage/commit files directly produced by current task
  • Never touch unrelated changes or other task outputs
  • Use git add <specific-files> instead of git add .
  • Verify staged files before commit to avoid cross-task conflicts

Multi-CLI Coexistence (CRITICAL):

  • If your task conflicts with existing uncommitted changes, STOP and report the conflict instead of overwriting
  • Treat all pre-existing uncommitted changes as intentional work-in-progress by other tools

System Optimization

Direct Binary Calls: Always call binaries directly in functions.shell, set workdir, avoid shell wrappers (bash -lc, cmd /c, etc.)

Text Editing Priority:

  1. Use apply_patch tool for all routine text edits
  2. Fall back to sed for single-line substitutions if unavailable
  3. Avoid Python editing scripts unless both fail

apply_patch invocation:

{
  "command": ["apply_patch", "*** Begin Patch\n*** Update File: path/to/file\n@@\n- old\n+ new\n*** End Patch\n"],
  "workdir": "<workdir>",
  "justification": "Brief reason"
}

Windows UTF-8 Encoding (before commands):

[Console]::InputEncoding  = [Text.UTF8Encoding]::new($false)
[Console]::OutputEncoding = [Text.UTF8Encoding]::new($false)
chcp 65001 > $null

Context Acquisition (MCP Tools Priority)

For task context gathering and analysis, ALWAYS prefer MCP tools:

  1. mcp__ace-tool__search_context - HIGHEST PRIORITY for code discovery
    • Semantic search with real-time codebase index
    • Use for: finding implementations, understanding architecture, locating patterns
    • Example: mcp__ace-tool__search_context(project_root_path="/path", query="authentication logic")
  2. smart_search - Fallback for structured search
    • Use smart_search(query="...") for keyword/regex search
    • Use smart_search(action="find_files", pattern="*.ts") for file discovery
    • Supports modes: auto, hybrid, exact, ripgrep
  3. read_file - Batch file reading
    • Read multiple files in parallel: read_file(path="file1.ts"), read_file(path="file2.ts")
    • Supports glob patterns: read_file(path="src/**/*.config.ts")

Priority Order:

ACE search_context (semantic) → smart_search (structured) → read_file (batch read) → shell commands (fallback)

NEVER use shell commands (cat, find, grep) when MCP tools are available.

Workflow Session Awareness

WorkflowDirectorySummary File
workflow-plan.workflow/active/WFS-*/workflow-session.json
workflow-lite-plan.workflow/.lite-plan/{date}-{slug}/plan.json
analyze-with-file.workflow/.analysis/ANL-*/conclusions.json
multi-cli-plan.workflow/.multi-cli-plan/*/session-state.json
lite-fix.workflow/.lite-fix/*/fix-plan.json
Other.workflow/.debug/, .workflow/.scratchpad/, .workflow/archives/—

Before starting work, scan recent sessions (7 days) to avoid conflicts and reuse prior work:

  • Overlapping file scope → warn, suggest referencing prior session
  • Complementary findings → feed into current task context

Execution Checklist

Before:

  • Understand PURPOSE and TASK clearly
  • Use ACE search_context first, fallback to smart_search for discovery
  • Use read_file to batch read context files, find 3+ patterns
  • Check RULES templates and constraints

During:

  • Follow existing patterns exactly
  • Write tests alongside code
  • Run tests after every change
  • Commit working code incrementally

After:

  • All tests pass
  • Coverage meets target
  • Build succeeds
  • All EXPECTED deliverables met

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.