agentleFS
HeimstattHeimstatt/living-skills/.cursor/rules/living-skills.mdc

living-skills / rules

Cursor rule0 starsChanged 43 days agoMITView on GitHub

Living Skills session rituals

  • Commits and pushes
  • Where files are headingNew Cursor rule files have more rules about planning and tracking tasks, and fewer about code style.

    Of the Cursor rule files we read, the 4,243 first seen in the last six months, against the older ones: the kind of rule that gained most share, and the kind that lost most. About files like this in general, not this one, and an estimate, since a kind is a pattern in the words.

  • In this repository2 other files it gives its agents.

The file

We split this living-skills.mdc into the rules it gives an agent

8 rules. About 0 are common, 3 are distinctive, and 5 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 5

    Very specific

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

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

    Read the relevant SKILL.md — understand the approach

    line 37
  2. Very specific:

    Read the relevant living-checklist.md — load accumulated knowledge

    line 38
  3. Distinctive:

    Output the Skill Activation Protocol: before beginning work:

    line 39
  4. Distinctive:

    After completing a task that used a Living Skill:

    line 55
  5. Very specific:

    Write new learnings to living-checklist.md:

    line 57
  6. Very specific:

    If skills were added, renamed or removed, refresh the skill mapping (symlinks under ~/.claude/skills/, see setup/skill-mapping.md):

    line 72
  7. Very specific:

    Write only to your own Team Memory/<instance>/ folder.

    line 86
  8. Distinctive:

    Update this list as you add skills:

    line 92

≈ “Common” is an estimate: at least 1 in 5 of the 6,809 Cursor rule 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
living-skills.mdc
---
description: Living Skills session rituals
alwaysApply: true
---
# Living Skills — Team Configuration

> This file is the single source of truth for all AI tools in this repository.
> `CLAUDE.md` (Claude Code) and `AGENTS.md` (Codex) are symlinks pointing here.
> `.cursor/rules/living-skills.mdc` is generated from this file — run `bash scripts/generate-cursor-rules.sh` after edits
> (Cursor: not actively tested in recent releases; generator kept for reference).
> Edit only this file — all tools pick up the changes automatically.

## Who I am

Set your instance name here. It appears in every commit message:
```
<Your-Instance-Name>: <what changed> [YYYY-MM-DD]
```

Examples: `Claude-Desktop`, `Claude-Server`, `Claude-Laptop`

## Repository layout

```
Infrastructure/skills/     ← Domain Skills for systems and tools
Projects/<name>/skills/    ← Domain Skills for specific projects
Team Memory/skills/        ← Behavioral Skills (methodologies, workflows)
Team Memory/<instance>/    ← Your memory: status.md, config.md
Team Memory/shared/        ← Human-contributed context
```

## Session Start

At the beginning of every session involving a Living Skill:

1. `git pull` — get the latest learnings from all instances
2. Read the relevant `SKILL.md` — understand the approach
3. Read the relevant `living-checklist.md` — load accumulated knowledge
4. **Output the Skill Activation Protocol** before beginning work:
   ```
   SKILL ACTIVATED: <skill-name>
   Date: YYYY-MM-DD
   Checklist read: Yes — [N] active entries, newest: [date of most recent entry]
   Active rule: "[verbatim quote of the most recent relevant rule]"
   Approach: [2–3 sentences on what this skill will do in this session]
   ```
   Domain skills add one more line, after `Checklist read:`:
   ```
   Knowledge checked: <date>, against <system/version>
   ```
   State it, do not judge it. Behavioral skills omit this line.

## Session End

After completing a task that used a Living Skill:

1. Write new learnings to `living-checklist.md`:
   ```
   ### [YYYY-MM-DD] — [task context] · *unconfirmed (1 observation)*
   **Learning:** what was discovered or what failed
   **Why it matters:** context and consequences
   **Rule:** a concrete guideline for next time
   **Observed by:** <your instance> (YYYY-MM-DD)
   ```
   The `unconfirmed` marker and `Observed by:` line are **optional**: many teams
   verify before writing, through review and counter-tests, rather than
   confirming later. If you use them and hit a pattern another instance already
   recorded, do not open a second entry — remove its marker and add yourself to
   `Observed by:`.
   Once a checklist has more than 50 entries, keep a `## Distilled Rules` section
   at its top.
2. If skills were added, renamed or removed, refresh the skill mapping
   (symlinks under `~/.claude/skills/`, see `setup/skill-mapping.md`):
   ```bash
   bash scripts/generate-claude-skills.sh
   ```
3. Commit and push:
   ```bash
   git add <skill-path>/living-checklist.md
   git commit -m "<Your-Instance-Name>: <what changed> [YYYY-MM-DD]"
   git push
   ```

## Write boundaries

Write only to your own `Team Memory/<instance>/` folder.
Other instances' folders are read-only — even if a task seems to call for it.
Each instance documents its own work so others can read it.

## Available skills

Update this list as you add skills:

| Skill | Type | Path | Use for |
|-------|------|------|---------|
| **token-optimization** | behavioral | `Team Memory/skills/token-optimization/` | **Proactively before any non-trivial task** — context budget, knowledge reuse, sub-agent strategy, caveman output compression |
| **surgical-changes** | behavioral | `Team Memory/skills/surgical-changes/` | **Any change to existing code** (≥2 files / 20 lines) — scope contract before the edit, R/O/U diff audit after, noticed-list instead of drive-by fixes. Not for typos or one-liners |
| **advocatus-diaboli** | behavioral | `Team Memory/skills/advocatus-diaboli/` | Critical analysis, strategy review, stress-testing assumptions |
| **secure-architecture** | behavioral | `Team Memory/skills/secure-architecture/` | Auth/authorization model reviews: AuthN≠AuthZ, stable-ID anchors, SoD in the data model, protecting authorization records |
| **business-analysis** | behavioral | `Team Memory/skills/business-analysis/` | BABOK/CBAP procedure: stakeholders → personas → user stories with AC → FR/NFR → gap analysis → risk register → out-of-scope |
| **skill-creator** | behavioral | `Team Memory/skills/skill-creator/` | Creating new Living Skills |
| **pair-review** | behavioral | `Team Memory/skills/pair-review/` | Cross-model code review: findings, critical response, re-review against a named commit |
| **strategy** | behavioral | `Team Memory/skills/strategy/` | Strategy staff unit: diagnose the question, route to a role (`roles/`), return advice to the owning department |
| **home-assistant** | domain | `Infrastructure/skills/home-assistant/` | HA automations, entities, integrations, debugging |
| **playwright** | domain | `Infrastructure/skills/playwright/` | Browser automation and visual UI checks with playwright-cli |
| **siyuan** | domain | `Infrastructure/skills/siyuan/` | **Deprecated** — optional example of a human-readable layer; use Obsidian instead ([`setup/obsidian-sync.md`](setup/obsidian-sync.md)) |
| *(browse)* | domain | `Infrastructure/skills/` | Other domain skills — systems and tools |

**Token optimization — proactive rule:** Activate `token-optimization` **without being asked** whenever a task will likely consume >10K tokens, involve sub-agents, or use multiple MCP tools. Exempt: short answers, single-turn Q&A.

## First time setup?

See [`setup/onboarding-new-instance.md`](setup/onboarding-new-instance.md).

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.