agentleFS
gyujeongiongyujeongion/claude-life-agent/CLAUDE.md

claude-life-agent

CLAUDE.md2 starsChanged 4 months agoMITView on GitHub

This is the master entry point for your personal AI workspace. Claude Code reads this file first, then routes to the right folder. Fill in the {{PLACEHOLDER}} tokens once, then let the structure work. 이 파일은 개인 AI 워크스페이스의 *마스터 진입점*입니다. 클로드 코드가 가장 먼저 읽고, 여기서 알맞은 폴더로 라우팅합니다. {{PLACEHOLDER}}만 한 번 채우면 됩니다. Owner: {{OWNER_NAME}} | Language: {{PRIMARY_LANGUAGE}} | Timezone: {{TIMEZONE}} ⚙️ FIRST-RUN TRIGGER (agent, read this first). If any {{PLACEHOLDER}} token still appears above or anywhere in…

The file

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

7 rules. About 0 are common, 4 are distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈4

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 3

    Very specific

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

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

    Every plan/proposal/draft starts its first line with one of these:

    line 96
  2. Distinctive:

    When the owner rejects a proposal and gives a reason:

    line 125
  3. Distinctive:

    only patterns seen 2+ times (never one-offs)

    line 140
  4. Distinctive:

    after logging, tell the owner in one line

    line 142
  5. Distinctive:

    Keep analysis docs and proposal docs separate.

    line 152
  6. Very specific:

    When revising, preserve the old version (_v1, _v2).

    line 153
  7. Very specific:

    Read docs/ARCHITECTURE.md for why it's built this way.

    line 164

≈ “Common” is an estimate: at least 1 in 5 of the 27,641 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 — Life Agent Workspace (master router)

This is the master entry point for your personal AI workspace. Claude Code reads this file first, then routes to the right folder. Fill in the {{PLACEHOLDER}} tokens once, then let the structure work.

이 파일은 개인 AI 워크스페이스의 마스터 진입점입니다. 클로드 코드가 가장 먼저 읽고, 여기서 알맞은 폴더로 라우팅합니다. {{PLACEHOLDER}}만 한 번 채우면 됩니다.

Owner: {{OWNER_NAME}} | Language: {{PRIMARY_LANGUAGE}} | Timezone: {{TIMEZONE}}


⚙️ FIRST-RUN TRIGGER (agent, read this first). If any {{PLACEHOLDER}} token still appears above or anywhere in this workspace — or the user asks you to "set this up" — stop and run ONBOARDING.md now, before answering anything else. It's an interview script that tailors this generic template to the user's actual life and domains. Do not just fill in a name; follow the whole script so the workspace becomes genuinely theirs.

⚙️ 최초 실행 트리거. 위쪽이나 워크스페이스 어디든 {{PLACEHOLDER}}가 남아있거나 사용자가 "세팅해줘"라고 하면 — 다른 답변보다 먼저 ONBOARDING.md를 실행하세요. 이름만 채우지 말고 스크립트 전체를 따라 사용자 삶·도메인에 맞게 재구성하세요.


The idea in one paragraph

This workspace is your life, organized as concentric circles so an AI agent can navigate it the way you would. The center (00_SELF) is who you are. The rings move outward: active work, ongoing roles, knowledge, systems, personal life. Every folder has its own CLAUDE.md that tells the agent what lives there and when to open it. The agent never scans everything — it follows pointers, exactly like a human opening the one drawer they need.


Session start — file loading strategy

Tier 1 — always loaded (auto-injected every session)

1. 00_SELF/memory/shared_memory.md   ← recent cross-session context
2. 00_SELF/memory/todo.md            ← only mention items due today/tomorrow

Wire these into a SessionStart hook so they load automatically. See docs/SETUP.md. Drift guard: if the hook is not actually wired (or you can't verify it), Read these two files yourself at session start. A doc that claims auto-injection while nothing is wired is the most common way this tier silently dies.

Tier 2 — on demand (loaded only when the request type calls for it)

3. 00_SELF/decisions/decisions_log.md       ← strategy / planning requests
4. 00_SELF/identity/character.md            ← taste / preference judgments
5. 00_SELF/memory/projects/INDEX.md         ← when a project keyword is detected
   → match trigger_keywords, then load only that projects/{slug}.md
   → rules: 00_SELF/memory/MEMORY_RULES.md

Navigation protocol

request arrives
  → pass the 00_SELF/context_router.md gate (classify context, decide which principles apply)
  → use the mapping table below to pick the folder
  → read that folder's CLAUDE.md
  → open ONLY the specific files you need

Never: list the whole root, wander into unrelated folders, print secrets, open files "just for reference."


Request → folder mapping

Request typeFolderEntry file
Who you are · identity · taste · memory00_SELF/00_SELF/CLAUDE.md
People · relationships00_SELF/people/people.md
Confirmed decisions · rejection history00_SELF/decisions/decisions_log.md
Active projects (have a deadline)10_WORK/10_WORK/CLAUDE.md
Ongoing roles / domains (no deadline)20_DOMAINS/20_DOMAINS/CLAUDE.md
Research · analysis · knowledge30_KNOWLEDGE/30_KNOWLEDGE/CLAUDE.md
Code · automation · infrastructure40_SYSTEMS/40_SYSTEMS/CLAUDE.md
Money · family · health · personal life50_LIFE/50_LIFE/CLAUDE.md
Done · inactive80_ARCHIVE/—
Shared / companion agent (isolated)90_SHARED/90_SHARED/CLAUDE.md
A note/memo/record request→ identify the domain above, save into that docthat domain's folder
Unclear—ask the owner directly

Document status tags

Every plan/proposal/draft starts its first line with one of these:

status: draft | active | on-hold | abandoned | complete
TagMeaningHow the agent treats it
draftbeing writtenincomplete — do not rely on
activein progressuse as current truth
on-holdpausedreference OK, do not execute
abandonedrejectednever carry into future plans (reason → feedback_patterns.md)
completedonecandidate to move into 80_ARCHIVE/

Source separation

TagMeaningWhere it lives
[OWNER]the owner said/confirmed it00_SELF/decisions/
[AI]the agent inferred/proposed it30_KNOWLEDGE/

On conflict, [OWNER] always wins.


Rejection-capture protocol

When the owner rejects a proposal and gives a reason:

  1. Mark that document status: abandoned.
  2. Classify the reason:
    • taste / disposition → 00_SELF/decisions/feedback_patterns.md
    • direction / brand → the relevant identity file
    • a changed decision → 00_SELF/decisions/decisions_log.md
  3. The next time you plan, that pattern is filtered automatically.

AI self-observation (optional but powerful)

The agent may, on its own, log repeated behavioral patterns it notices in the owner to 00_SELF/identity/character_ai_mirror.md. Rules:

  • only patterns seen 2+ times (never one-offs)
  • descriptive, not evaluative ("tends to…", never "is bad at…")
  • after logging, tell the owner in one line

When you plan or strategize (checklist)

  1. 00_SELF/memory/SYNC_AXIS.md — does this align with the owner's core axis?
  2. 00_SELF/decisions/decisions_log.md — any conflict with a confirmed decision?
  3. 00_SELF/decisions/feedback_patterns.md — apply past rejection filters.
  4. 00_SELF/identity/character.md — does this fit the owner's disposition?
  5. Keep analysis docs and proposal docs separate.
  6. When revising, preserve the old version (_v1, _v2).

The two gates (this is the whole trick)

  1. Navigation gate (this file): decides which files to open.
  2. Answer gate (00_SELF/context_router.md): decides which principles to apply to the answer — because advice about your career should not be governed by the same rules as advice about your health or your relationships.

Read docs/ARCHITECTURE.md for why it's built this way.

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.