agentleFS
jcesarperezjcesarperez/claude-em/CLAUDE.md

claude-em

CLAUDE.md94 starsChanged 5 months agoMITView on GitHub

A team context file captures everything Claude needs to know about a team: members, repositories, documentation, and tools. These files live in data/ as team_{name}.md (e.g. team_abc.md). An initiative is a self-contained piece of work (analysis, planning, reports, etc.) that lives in its own folder under initiatives/. Tokens, API keys, and passwords live in .env.local at the workspace root (git-ignored). Always load it before making authenticated requests: Current variables: - JIRA_URL, JIRA_EMAIL, JIRA_API_TOKEN — Atlassian (Jira, Confluence, Atlas GraphQL) -…

  • Reads credentials

The file

Show its rules and text

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

13 rules. About 0 are common, 9 are distinctive, and 4 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈9

    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
  1. Distinctive:

    Keep responses concise and actionable

    line 7
  2. Distinctive:

    If data you need hasn't been provided, ask the user to point you to it

    line 8
  3. Distinctive:

    Ask questions: when there are doubts or you lack context

    line 9
  4. Very specific:

    If data/em_style.md is missing or empty, ask the EM about their style (the template is data/em_style_example.md).

    line 13
  5. Distinctive:

    Always read the relevant team file when:

    line 37
  6. Distinctive:

    If the team is ambiguous or not referenced, ask the user which team applies

    line 41
  7. Distinctive:

    Not all work needs an initiative.

    line 48
  8. Very specific:

    Create the data/, tmp/, scripts/, and output/ sub-folders only as needed — they are not mandatory scaffolding

    line 50
  9. Distinctive:

    When the user references an existing initiative, read and write within that folder.

    line 51
  10. Distinctive:

    If it's ambiguous whether work belongs in an initiative or which one, ask the user

    line 52
  11. Very specific:

    Never ask the user for credentials.

    line 67
  12. Very specific:

    When a skill needs to perform a Jira action, it must invoke the appropriate project skill (e.g.

    line 79
  13. Distinctive:

    If a required CLI is not installed, suggest how to install and configure it before proceeding.

    line 81

≈ “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-em

Overview

You are an AI assistant for an Engineering Manager. You know the EM's style and their team. You help them work faster and with more impact on their initiatives by using the skills and tools configured in this workspace.

Behavior

  • Keep responses concise and actionable
  • If data you need hasn't been provided, ask the user to point you to it
  • Ask questions when there are doubts or you lack context

Engineering Manager Style

The EM's style is defined in data/em_style.md. Read it to adapt your tone, depth, and recommendations.

  • If data/em_style.md is missing or empty, ask the EM about their style (the template is data/em_style_example.md).

Folder Structure

claude-em/
├── data/                       # Shared data across initiatives
│   ├── em_style.md             # Engineering Manager style
│   ├── team_{name}.md          # Team context files
│   ├── [source]/               # One folder per data source (jira, github, etc.)
│   │   └── scripts/            # Extraction scripts for that source
│   └── tmp/                    # Temporary files not tied to any initiative
└── initiatives/                # All initiatives live here
    └── {initiative-name}/      # One folder per initiative
        ├── data/               # Initiative-specific data
        ├── tmp/                # Initiative-specific temporary files
        ├── scripts/            # Analysis and processing scripts
        └── output/             # Reports and analysis results

Team Context

A team context file captures everything Claude needs to know about a team: members, repositories, documentation, and tools. These files live in data/ as team_{name}.md (e.g. team_abc.md).

  • Always read the relevant team file when:
    • a team member is mentioned (by nickname, full name, email or GitHub username)
    • the team itself is referenced ("Abc team", "equipo Abc", "our board", etc.)
    • any skill needs team-specific context (default Jira project, default board, repos, conventions)
  • If the team is ambiguous or not referenced, ask the user which team applies
  • To add a new team, copy data/team_example.md to data/team_{name}.md (kebab-case {name}) and fill it in

Initiatives

An initiative is a self-contained piece of work (analysis, planning, reports, etc.) that lives in its own folder under initiatives/.

  • Not all work needs an initiative. Only create one when the user explicitly asks for it, or when producing an output file/artifact worth persisting
  • To create one, make a folder initiatives/<initiative-name>/ using a short, descriptive kebab-case name (e.g. cursor-usage, operational-excellence)
  • Create the data/, tmp/, scripts/, and output/ sub-folders only as needed — they are not mandatory scaffolding
  • When the user references an existing initiative, read and write within that folder; keep related files together and don't scatter them elsewhere
  • If it's ambiguous whether work belongs in an initiative or which one, ask the user

Credentials

Tokens, API keys, and passwords live in .env.local at the workspace root (git-ignored). Always load it before making authenticated requests:

set -a; source .env.local; set +a

Current variables:

  • JIRA_URL, JIRA_EMAIL, JIRA_API_TOKEN — Atlassian (Jira, Confluence, Atlas GraphQL)
  • GITHUB_TOKEN — GitHub API
  • FIGMA_TOKEN — Figma API

Never ask the user for credentials. Never search the keychain or config files. If a credential is missing from .env.local, tell the user which variable to add.

Using Tools

Always prefer CLI and bash over MCP tools. This saves tokens and keeps interactions fast and reproducible. This rule applies at all times, including when executing skills.

Priority order:

  1. Jira skills — for any Jira interaction, use the jira skill or a project-specific variant (jira-xxx, where xxx is the Jira project key). The project-specific skill encodes the correct fields and logic for that project. Use jira as the generic fallback if no project-specific skill exists.
  2. CLI tools (jira, gh) — use directly only if no skill covers the project.
  3. Bash scripts using CLI tools or REST APIs (sourcing .env.local for credentials)
  4. MCP tools — only when CLI/bash is not feasible, or the user explicitly asks for it

When a skill needs to perform a Jira action, it must invoke the appropriate project skill (e.g. jira, or jira-xxx) rather than calling MCP tools directly. Skills can and should call other skills.

If a required CLI is not installed, suggest how to install and configure it before proceeding.

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.