docs: add LLM Reference section and Learning Mode documentation
- Add LLM Reference section in README with curl one-liner for AI assistants - Add learning mode flags and deep_dive refs to YAML index - Add comprehensive Learning Mode documentation to guide (installation, usage, integration) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
This commit is contained in:
Florian BRUNIAUX
parent
cba3003116
commit
471e1a0d7a
3 changed files with 155 additions and 0 deletions
13
README.md
13
README.md
|
|
@ -64,6 +64,7 @@ claude
|
|||
| Need to configure a team setup | [Power User Path](#-by-role-tailored-learning-paths) | 2h |
|
||||
| Need to evaluate/approve adoption | [PM Path](#-by-role-tailored-learning-paths) | 20 min |
|
||||
| Want to check your current setup | [Audit Your Setup](#-audit-your-setup) | 2 sec |
|
||||
| Want AI assistants to know Claude Code | [LLM Reference](#-llm-reference) | 1 curl |
|
||||
|
||||
### ⚡ Audit Your Setup
|
||||
|
||||
|
|
@ -77,6 +78,18 @@ curl -sL https://raw.githubusercontent.com/FlorianBruniaux/claude-code-ultimate-
|
|||
|
||||
**Want deeper analysis?** Use [`claude-setup-audit-prompt.md`](./claude-setup-audit-prompt.md) for personalized recommendations (~3 min).
|
||||
|
||||
### 🤖 LLM Reference
|
||||
|
||||
Give any AI assistant instant Claude Code expertise (~2K tokens):
|
||||
|
||||
```bash
|
||||
curl -sL https://raw.githubusercontent.com/FlorianBruniaux/claude-code-ultimate-guide/main/claude-code-reference.yaml
|
||||
```
|
||||
|
||||
**Use cases**: Paste output into ChatGPT/Claude/Gemini, add to system prompts, or reference in Claude Code with `@claude-code-reference.yaml`
|
||||
|
||||
**What's inside**: Decision trees, command reference, context zones, MCP servers, agent templates, troubleshooting—optimized for machine consumption. Points to line numbers in the [full guide](./english-ultimate-claude-code-guide.md) for deep dives.
|
||||
|
||||
### 🎯 By Role (Tailored Learning Paths)
|
||||
|
||||
<table>
|
||||
|
|
|
|||
|
|
@ -51,6 +51,8 @@ deep_dive:
|
|||
troubleshooting: 7892
|
||||
cheatsheet: 8265
|
||||
daily_workflow: 8341
|
||||
superclaude_modes: 8634
|
||||
learning_mode: 8645
|
||||
|
||||
# ════════════════════════════════════════════════════════════════
|
||||
# DECISION TREE (most important - en premier)
|
||||
|
|
@ -62,6 +64,7 @@ decide:
|
|||
repeating: "create agent or command"
|
||||
need_docs: "Context7 MCP"
|
||||
deep_debug: "--think or --ultrathink"
|
||||
learning: "--learn (want to understand decisions)"
|
||||
|
||||
# ════════════════════════════════════════════════════════════════
|
||||
# PROMPTING FORMULA (see deep_dive.xml_prompting for details)
|
||||
|
|
@ -128,6 +131,9 @@ cli:
|
|||
"--dangerously-skip-permissions": "auto-accept ALL (danger)"
|
||||
"--debug": "verbose output"
|
||||
"--mcp-debug": "debug MCP servers"
|
||||
"--learn": "enable learning mode (explanations)"
|
||||
"--learn focus:X": "learn only for domain (git/arch/sec)"
|
||||
"--no-learn": "suppress learning offers"
|
||||
|
||||
# ════════════════════════════════════════════════════════════════
|
||||
# CONTEXT MANAGEMENT - see deep_dive.context_management
|
||||
|
|
|
|||
|
|
@ -8629,6 +8629,142 @@ SuperClaude transforms Claude Code into a structured development platform throug
|
|||
- Smart personas for different contexts
|
||||
- MCP server integration
|
||||
- Task management and session persistence
|
||||
- **Behavioral modes** for optimized workflows
|
||||
|
||||
#### SuperClaude Behavioral Modes
|
||||
|
||||
SuperClaude includes configurable behavioral modes stored in `~/.claude/MODE_*.md` files:
|
||||
|
||||
| Mode | Purpose | Activation |
|
||||
|------|---------|------------|
|
||||
| **Orchestration** | Smart tool selection, parallel execution | Auto (multi-tool ops, >75% context) |
|
||||
| **Task Management** | Hierarchical task tracking with memory | Auto (>3 steps, >2 directories) |
|
||||
| **Token Efficiency** | Symbol-enhanced compression (30-50% reduction) | Auto (>75% context) or `--uc` |
|
||||
| **Learning** | Just-in-time skill development | `--learn` flag or "why/how" questions |
|
||||
|
||||
#### Learning Mode: Installation & Usage
|
||||
|
||||
Learning Mode provides contextual explanations when techniques are first used, without overwhelming you with repeated explanations.
|
||||
|
||||
**Installation**:
|
||||
|
||||
1. Create the mode file:
|
||||
```bash
|
||||
# Create MODE_Learning.md in your global Claude config
|
||||
touch ~/.claude/MODE_Learning.md
|
||||
```
|
||||
|
||||
2. Add the content (or copy from SuperClaude framework):
|
||||
```markdown
|
||||
# Learning Mode
|
||||
|
||||
**Purpose**: Just-in-time skill development with contextual explanations when techniques are first used
|
||||
|
||||
## Activation Triggers
|
||||
- Manual flag: `--learn`, `--learn focus:[domain]`
|
||||
- User explicitly asks "why?" or "how?" about an action
|
||||
- First occurrence of advanced technique in session
|
||||
|
||||
## Default Behavior
|
||||
**OFF by default** - Activates via triggers above or explicit `--learn` flag
|
||||
|
||||
When active, tracks techniques explained this session to avoid repetition.
|
||||
```
|
||||
|
||||
3. Register in `~/.claude/CLAUDE.md`:
|
||||
```markdown
|
||||
# Behavioral Modes
|
||||
@MODE_Learning.md
|
||||
```
|
||||
|
||||
4. Add flags to `~/.claude/FLAGS.md`:
|
||||
```markdown
|
||||
**--learn**
|
||||
- Trigger: User requests learning mode, beginner signals, "why/how" questions
|
||||
- Behavior: Enable just-in-time explanations with first-occurrence tracking
|
||||
|
||||
**--no-learn**
|
||||
- Trigger: User wants pure execution without educational offers
|
||||
- Behavior: Suppress all learning mode offers
|
||||
```
|
||||
|
||||
**Usage**:
|
||||
|
||||
```bash
|
||||
# Activate for entire session
|
||||
claude --learn
|
||||
|
||||
# Focus on specific domain
|
||||
claude --learn focus:git
|
||||
claude --learn focus:architecture
|
||||
claude --learn focus:security
|
||||
|
||||
# Batch explanations at end
|
||||
claude --learn batch
|
||||
```
|
||||
|
||||
**Offer Format**:
|
||||
|
||||
When Learning Mode is active, Claude offers explanations after technical actions:
|
||||
|
||||
```
|
||||
git rebase -i HEAD~3
|
||||
-> Explain: rebase vs merge? (y/detail/skip)
|
||||
```
|
||||
|
||||
Response options:
|
||||
- `y` → Surface explanation (20-50 tokens)
|
||||
- `detail` → Medium depth (100-200 tokens)
|
||||
- `skip` → Continue without explanation
|
||||
|
||||
**With Token Efficiency Mode** (compressed format):
|
||||
```
|
||||
git rebase -i HEAD~3
|
||||
-> ?rebase
|
||||
```
|
||||
|
||||
**Integration with Other Modes**:
|
||||
|
||||
| Combined With | Behavior |
|
||||
|---------------|----------|
|
||||
| Token Efficiency (`--uc`) | Compressed offer format: `-> ?[concept]` |
|
||||
| Task Management | Batch explanations at phase completion |
|
||||
| Brutal Advisor | Brutal on diagnosis, pedagogical on explanation |
|
||||
|
||||
**Priority Rules**:
|
||||
```
|
||||
--no-learn > --uc > --learn
|
||||
Token Efficiency constraints > Learning verbosity
|
||||
Task flow > Individual explanations
|
||||
```
|
||||
|
||||
**Example Session**:
|
||||
|
||||
```bash
|
||||
$ claude --learn
|
||||
|
||||
You: Refactor the authentication module
|
||||
|
||||
Claude: [Reads files, implements changes]
|
||||
git rebase -i HEAD~3
|
||||
-> Explain: rebase vs merge? (y/detail/skip)
|
||||
|
||||
You: y
|
||||
|
||||
Claude: Rebase rewrites history linearly; merge preserves branches.
|
||||
Use rebase for clean history before push, merge for shared branches.
|
||||
|
||||
[Continues work - won't ask about rebase again this session]
|
||||
```
|
||||
|
||||
**When to Use Learning Mode**:
|
||||
|
||||
| Use `--learn` | Use `--no-learn` |
|
||||
|---------------|------------------|
|
||||
| New to a technology | Expert in the domain |
|
||||
| Onboarding to project | Time-critical tasks |
|
||||
| Want to understand decisions | Already know the patterns |
|
||||
| Mentoring yourself | High context pressure |
|
||||
|
||||
#### Learning Sites
|
||||
|
||||
|
|
|
|||
Loading…
Add table
Add a link
Reference in a new issue