agentleFS
joris887joris887/exosuit/CLAUDE.md

exosuit

CLAUDE.md9 starsChanged 3 months agoMITView on GitHub

Profile: standard <!-- Project complexity profile: lean | standard | strict. Set by /bootstrap or manually. --> <!-- Override per-session: EXOSUIT_PROJECT_PROFILE=lean|standard|strict -->

The file

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

7 rules. About 0 are common, 6 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈6

    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. Distinctive:

    NEVER hardcode secrets — use environment variables

    line 18
  2. Very specific:

    NEVER use any types without justification comment

    line 19
  3. Distinctive:

    NEVER weaken test assertions to make tests pass -->

    line 20
  4. Distinctive:

    All API endpoints use REST, not GraphQL (ADR-0003)

    line 41
  5. Distinctive:

    Before proposing any architectural change, check @docs/adr/ for prior decisions.

    line 44
  6. Distinctive:

    When compacting context, use this structured key-value format.

    line 122
  7. Distinctive:

    When multiple compactions occur:

    line 152

≈ “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.

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

[Project Name]

Project Overview

<!-- Filled by /bootstrap -->

Profile: standard

<!-- Project complexity profile: lean | standard | strict. Set by /bootstrap or manually. --><!-- Override per-session: EXOSUIT_PROJECT_PROFILE=lean|standard|strict -->

Tech Stack

<!-- /bootstrap: Fill with detected versions. Prevents AI from mixing incompatible API versions. --><!-- Example: Python 3.12 / FastAPI 0.109 / SQLAlchemy 2.0 / Pydantic 2.5 -->

Critical Rules

<!-- /bootstrap: Promote the 3-5 most project-damaging violations from CODING_STANDARDS.md --><!-- These load every session. Only rules where violation causes real damage belong here. --><!-- Example: - NEVER hardcode secrets — use environment variables - NEVER use `any` types without justification comment - NEVER weaken test assertions to make tests pass -->

Commands

# Filled by /bootstrap with detected commands:
# dev:       <dev server command>
# test:      <test command>
# lint:      <lint command>
# format:    <format command>
# build:     <build command>
# typecheck: <typecheck command>

Architecture

<!-- One-liner filled by /bootstrap. See docs/architecture/ARCHITECTURE.md for details. -->

Architecture Rules

<!-- Derived from accepted ADRs. See the linked ADR for full rationale. --><!-- /bootstrap: Populate from existing docs/adr/ records. Add manually as ADRs are accepted. --><!-- Example: - All API endpoints use REST, not GraphQL (ADR-0003) - Authentication uses Auth0; do not implement custom auth (ADR-0015) -->
  • Before proposing any architectural change, check @docs/adr/ for prior decisions.

Git Workflow

  • Default branch: <!-- Detected by /bootstrap (main, master, develop, etc.) -->
  • Flow: Modified GitHub Flow (sprint branches → squash merge to main)
  • Branches: <type>/<description> (feat, fix, hotfix, refactor, docs, test, chore) or sprint-<N>
  • Commits: Conventional Commits: <type>(<scope>): <description> — atomic, one logical unit each
  • PRs: ≤400 LOC (target ≤200). Squash merge to main. Delete branch after merge.
  • Safety: Force push, reset --hard, clean -f, branch -D, --no-verify, checkout . — all blocked by hooks
  • Rollback: /rewind → reset --soft → stash → revert → (reset --hard blocked)
  • Reference: See docs/reference/GIT_WORKFLOW.md for full guide

Current Focus

<!-- Updated per sprint by /sprint-end -->

Backlog

  • Index: @docs/reference/BACKLOG_INDEX.md
  • Epics: @docs/reference/backlog/E##-name.md (load only current epic)
  • Progress: @docs/progress.md

Testing

TDD mandatory for feature, bug fix, and refactoring stories. See docs/reference/TESTING_STRATEGY.md.

Skills — Exosuit v5.0

See .claude/skills/SKILLS_INVENTORY.md for full inventory.

Core Workflow

/sprint-start → /story-cycle (repeat) → /sprint-end
I want to...Use
Setup framework/bootstrap
Resume work/continue
Start sprint/sprint-start
Deliver story/story-cycle <description>
Plan work/ideate <idea>
Discover project/discover "<idea>"
Explore design/brainstorm <idea>
Deep research/research <topic>
End sprint/sprint-end
End session/handoff
Leave safely/wrap-up
Debug an issue/debug-session <error>
Test plan/manual-test
Test feedback/testing-cycle <feedback>
UAT test case/UAT-cycle <test-case-id>
Sense check UAT cases/claude-sense-check
Parallel work/parallel-work
Publish stream work/merge-up
Refresh stream/merge-down
Parallel work in tmux/lanes start a b
Architecture check/architecture-check
Framework health check/doctor
Upgrade framework/framework-upgrade
Evaluate a skill/skill-eval <mode> <skill>
Phase review/phase-review [phase]
Undo failed work/undo-work
Run quality gates/quality-check [--all]
Iterative refinement/refine-loop "<task>"
Optimize a metric/optimize "<goal>"
<!-- Profile: lean — /bootstrap generates only these ~12 core skills in the table: /bootstrap, /continue, /story-cycle, /ideate, /sprint-start, /sprint-end, /handoff, /debug-session, /research, /doctor, /help-me, /quality-check -->

Important Files

  • docs/reference/CODING_STANDARDS.md — Code conventions
  • docs/reference/TESTING_STRATEGY.md — Testing practices
  • docs/reference/STORY_SIZING.md — Project sizing policy (sizing is by cohesion, not file count)
  • docs/architecture/ARCHITECTURE.md — System architecture
  • docs/reference/GROUND_RULES.md — Architectural principles
  • docs/adr/ — Architecture decision records (prior decisions and rejected alternatives)
  • docs/reference/MCP_INTEGRATION.md — MCP server selection and integration guide
  • docs/testing/UAT_COVERAGE.md — UAT test cases

Compaction Directive

When compacting context, use this structured key-value format. Items are tagged by priority — drop LOW first, CRITICAL must survive all compactions verbatim.

## Compacted Context

### [CRITICAL — preserve verbatim across all compactions]
goal: "[Current objective — what story/task is in progress]"
sprint_goal: "[Current sprint goal — one sentence from sprint spec]"
sprint_spec: "docs/sprints/sprint-[N].md"
commands: {test: "[cmd]", lint: "[cmd]", build: "[cmd]", typecheck: "[cmd]"}
active_plan: |
  [Any in-progress plan, verbatim if possible]

### [HIGH — preserve if space allows]
branch: "[name]"
sprint: [number]
phase: "[current phase or status]"
decisions:
  - choice: "[decision]" | reason: "[rationale]"
in_progress: "[current task and its state]"

### [NORMAL — summarize if needed]
completed: ["[item1]", "[item2]"]
blocked_by: "[blocker or null]"

### [LOW — drop first, these are recoverable]
<files-read>[paths of all files explored this session, one per line]</files-read>
<files-modified>[paths of all files changed this session, one per line]</files-modified>

When multiple compactions occur:

  • MERGE new file paths into existing <files-read> and <files-modified> lists — never discard previous entries
  • Drop LOW items first if space is tight (files can be re-read)
  • CRITICAL items must survive every compaction — never summarize or drop them

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.