docs: update guide content, examples, tools, and reference files

- guide/ultimate-guide.md — content updates - guide/workflows/README.md, guide/README.md — navigation improvements - guide/diagrams/ — diagram updates (context/sessions, config, MCP ecosystem) - guide/third-party-tools.md — additions - examples/README.md, hooks/README.md, scripts/README.md — examples updates - examples/skills/pr-triage/SKILL.md — expanded skill - machine-readable/reference.yaml — reference sync - tools/audit-prompt.md, tools/onboarding-prompt.md — tooling updates - docs/for-cto.md, docs/for-tech-leads.md, docs/resource-evaluations/README.md — doc updates - .gitignore — gitignore update Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
2026-03-09 15:32:33 +01:00 · 2026-03-09 15:32:33 +01:00 · 8f1dcecfa2
commit 8f1dcecfa2
parent 9c02214bb8
19 changed files with 639 additions and 104 deletions
--- a/tools/audit-prompt.md
+++ b/tools/audit-prompt.md
@ -53,7 +53,7 @@ cd your-project-directory
 claude
 ```

-> **Note**: With Opus 4.5, thinking mode is enabled by default at maximum depth. Use Alt+T to toggle if needed.
+> **Note**: With Opus 4.6, thinking mode is enabled by default at maximum depth. Use Alt+T to toggle if needed.

 ### Step 3: Paste and Execute

@ -308,7 +308,7 @@ For each category, evaluate against these criteria based on Phase 1 scan results
 - [ ] Other relevant MCPs for the project needs

 **Thinking Mode & Trinity (Guide Section 9.1)**
- [ ] Understanding of thinking mode (enabled by default in Opus 4.5, Alt+T to toggle)
+- [ ] Understanding of thinking mode (enabled by default in Opus 4.6, Alt+T to toggle)
 - [ ] Trinity pattern documented for complex workflows

 **CI/CD Integration (Guide Section 9.3)**
@ -345,6 +345,26 @@ For each category, evaluate against these criteria based on Phase 1 scan results
 - [ ] Finding domain experts via git history
 - [ ] Understanding code evolution patterns

+**Rules Templates (Guide Section 3.2)**
+- [ ] `.claude/rules/` directory exists with auto-loaded rule files
+- [ ] Rules cover relevant domains (architecture, code quality, testing)
+- [ ] Rules are concise and actionable (not duplicating CLAUDE.md content)
+
+**Sandbox & Permissions (Guide Section 1.4)**
+- [ ] Understanding of sandbox modes (Docker container vs native process-level)
+- [ ] Permission modes configured appropriately for the project
+- [ ] Sensitive file patterns blocked via `permissions.deny` (`.env*`, `*.pem`, `credentials*`)
+
+**Security Commands (Guide: security-hardening.md)**
+- [ ] `/security-check` available for quick config scans (~30s)
+- [ ] `/security-audit` available for comprehensive audits (2-5min, scored /100)
+- [ ] Awareness of threat database (`threat-db.yaml`) for known attack patterns
+
+**Plan-Validate-Execute Pipeline (Guide: workflows/plan-pipeline.md)**
+- [ ] Awareness of pipeline workflow for complex features (power users)
+- [ ] `/plan-start`, `/plan-validate`, `/plan-execute` commands installed if applicable
+- [ ] ADR learning loop understood for accumulating architectural decisions
+
 #### 2.2 Calculate Health Score

 **Formula**: `Score = (earned_points / max_points) × 100`
@ -395,20 +415,23 @@ For each ❌ or ⚠️ item, provide:
 Instead of reading the entire guide, use these line ranges for targeted extraction:

 ```bash
-# Memory Files best practices
-sed -n '2184,2314p' guide/ultimate-guide.md
+# Line numbers from reference.yaml (deep_dive keys) — verify with:
+# grep "memory_files\|hooks\|mcp_servers\|context_management\|plan_mode" machine-readable/reference.yaml

-# Hooks section
-sed -n '3962,4528p' guide/ultimate-guide.md
+# Memory Files best practices (deep_dive: memory_files)
+sed -n '3054,3254p' guide/ultimate-guide.md

-# MCP Servers section
-sed -n '4529,5076p' guide/ultimate-guide.md
+# Hooks section (deep_dive: hooks)
+sed -n '8300,8800p' guide/ultimate-guide.md

-# Context Management
-sed -n '910,1423p' guide/ultimate-guide.md
+# MCP Servers section (deep_dive: mcp_servers)
+sed -n '9500,10000p' guide/ultimate-guide.md

-# Plan Mode
-sed -n '1424,1601p' guide/ultimate-guide.md
+# Context Management (deep_dive: context_management)
+sed -n '1000,1500p' guide/ultimate-guide.md
+
+# Plan Mode (deep_dive: plan_mode)
+sed -n '1500,1700p' guide/ultimate-guide.md
 ```

 **Suggested Templates**:
@ -509,12 +532,16 @@ Here's an example of what the audit report looks like:
 | **PreToolUse** | Hook that runs BEFORE Claude executes a tool - great for security checks |
 | **PostToolUse** | Hook that runs AFTER Claude executes a tool - great for formatting |
 | **Plan Mode** | Read-only exploration mode for safe analysis before making changes |
-| **Thinking Mode** | Extended thinking (Opus 4.5: ON by default). Toggle with Alt+T, configure in /config |
+| **Thinking Mode** | Extended thinking (Opus 4.6: ON by default). Toggle with Alt+T, configure in /config |
 | **Trinity Pattern** | Combining Plan Mode + Extended Thinking + MCP for complex tasks |
 | **Verify Gate** | CI/CD pattern: build → lint → test → typecheck before merge |
-| **Context Zones** | Green (0-50%), Yellow (50-70%), Red (70%+) - context usage thresholds |
+| **Context Zones** | < 70% optimal, 75% auto-compact trigger, 85% handoff recommended, 95% force handoff |
 | **Data Retention** | Anthropic stores conversations: 5 years (default), 30 days (opt-out), 0 days (Enterprise ZDR) |
 | **permissions.deny** | Settings to block Claude from reading sensitive files like `.env`, credentials |
+| **Rules** | Auto-loaded `.claude/rules/*.md` files providing contextual instructions every session |
+| **Permission Modes** | Trust levels for Claude's tool access: default deny, allowlist, or prompt-on-use |
+| **Sandbox** | OS-level isolation (Docker container or native process-level). Toggle with `/sandbox` |
+| **Plugins** | Community extensions installable via `/install-plugin owner/repo` |

 ### Priority Levels Explained

@ -585,4 +612,4 @@ Here's an example of what the audit report looks like:

 ---

-*Last updated: January 2026 | Version 2.9 - Fixed MCP detection (now checks ~/.claude.json)*
+*Last updated: March 2026 | Version 3.0 - Updated for guide v3.32.2 (Opus 4.6, new checklist categories, glossary, context zones)*
--- a/tools/onboarding-prompt.md
+++ b/tools/onboarding-prompt.md
@ -28,10 +28,10 @@ This prompt instructs Claude to become your personal onboarding coach by:

 | Goal | What You'll Get |
 |------|-----------------|
-| **Get started** | Golden Rules + essential commands + first workflow |
-| **Optimize** | Context management + Plan Mode + cost optimization |
-| **Build agents** | Agent/Skill/Command templates + hooks |
-| **Learn security** | Threat landscape + MCP vetting + scanning tools + hardening |
+| **Get started** | Golden Rules + sandbox modes + essential commands + first workflow |
+| **Optimize** | Context management + Plan Mode + cost optimization + Plan-Validate-Execute pipeline |
+| **Build agents** | Agent/Skill/Command templates + Skills 2.0 taxonomy + hooks |
+| **Learn security** | Sandbox modes + permission hardening + MCP vetting + scanning tools + threat DB |
 | **Fix a problem** | Direct jump to troubleshooting |
 | **Learn everything** | Complete guided tour |

@ -108,16 +108,21 @@ https://raw.githubusercontent.com/FlorianBruniaux/claude-code-ultimate-guide/mai

 **Adaptive topic selection (when reference.yaml loads successfully):**

-The onboarding matrix uses **adaptive architecture** (v2.0.0, guide v3.23.0+):
+The onboarding matrix uses **adaptive architecture** (v2.1.0, guide v3.32.2+):
 - Each profile has **core topics** (always shown) + **adaptive topics** (context-based)
- Claude analyzes user's initial messages for trigger keywords to surface relevant v3.21-3.22 content
+- Claude analyzes user's initial messages for trigger keywords to surface relevant v3.21-3.32 content
 - Keyword examples:
  - "team", "sync", "backup", "multi-machine" → `config_hierarchy` (backup/sync strategies)
  - "git", "version control", "commits" → `git_mcp_guide` (official Git MCP server)
  - "secrets", "API keys", "credentials" → `mcp_secrets_management` (secrets handling)
  - "quality", "review", "planner", "dual" → `dual_instance_planning` (planner/implementer pattern)
  - "security", "sandbox", "isolation" → `sandbox_native_guide` or `security_hardening`
- Ensures v3.21-3.22 features surface based on **relevance**, not just chronology
+  - "permission", "allow", "deny" → `permission_modes`
+  - "memory", "persist", "session" → `memory_files`
+  - "template", "structure", "format" → `skill_template`
+  - "validation", "checklist", "deploy" → `agent_validation_checklist`
+  - "plan", "pipeline" → `plan_pipeline_workflow` (Plan-Validate-Execute)
+- Ensures v3.21-3.32 features surface based on **relevance**, not just chronology
 - Respects time budgets (max 4-7 topics per profile, validated 6-8 min/topic)

 **Fallback if fetch fails:**
@ -127,7 +132,7 @@ If you cannot fetch the reference.yaml:
   - `get_started`: rules → sandbox_native_guide → commands
   - `optimize`: context_management → plan_mode → cost_optimization
   - `build_agents`: agents → skills → hooks
-   - `learn_security`: sandbox_native_guide → mcp_secrets_management → security_hardening
+   - `learn_security`: sandbox_native_guide → permission_modes → mcp_secrets_management → security_hardening
   - `fix_problem`: troubleshooting checklist
 3. Continue with Phase 1.5 questions as normal.

@ -156,6 +161,7 @@ Based on the goal from Phase 0, ask ONLY the necessary additional questions:
 - ⏱️ 15-30 min
 - 🎯 30-60 min
 - 📚 1+ hour
+- 📖 2+ hours

 **Style question** (if time >= 15min) - Use AskUserQuestion with options:
 - 📖 Explanations (tell me why)
@ -247,13 +253,13 @@ Based on time spent and topics covered:
   - `fix_problem` → "Run `claude doctor` if issues persist"

 3. **Next steps**: Point to relevant resources with clickable URLs:
-   - **Quiz (RECOMMENDED)** - Validate what you learned (274 questions total, 15 categories):
-     - Beginner (5min/15min/30min profiles): [Quiz - Basics (60 questions, ~15 min)](https://github.com/FlorianBruniaux/claude-code-ultimate-guide/tree/main/quiz#beginner-categories) - Categories: basics, commands, shortcuts, reference
-     - Intermediate (15min/30min profiles): [Quiz - Workflows (100 questions, ~25 min)](https://github.com/FlorianBruniaux/claude-code-ultimate-guide/tree/main/quiz#intermediate-categories) - Categories: workflows, context, agents, hooks
-     - Advanced/Power (30min/60min/120min profiles): [Quiz - Production (97 questions, ~30 min)](https://github.com/FlorianBruniaux/claude-code-ultimate-guide/tree/main/quiz#advanced-categories) - Categories: MCP, production, advanced, learning, ecosystem
-     - Security-focused: [Quiz - Security Hardening (30 questions)](https://github.com/FlorianBruniaux/claude-code-ultimate-guide-landing/tree/main/questions/13-security-hardening) - Attack techniques, CVEs, campaigns, scanning tools
+   - **Quiz (RECOMMENDED)** - Validate what you learned (271 questions, 15 categories):
+     - Beginner (5min/15min/30min profiles): [Quiz - Basics](https://github.com/FlorianBruniaux/claude-code-ultimate-guide/tree/main/quiz#beginner-categories) - Categories: basics, commands, shortcuts, reference
+     - Intermediate (15min/30min profiles): [Quiz - Workflows](https://github.com/FlorianBruniaux/claude-code-ultimate-guide/tree/main/quiz#intermediate-categories) - Categories: workflows, context, agents, hooks
+     - Advanced/Power (30min/60min/120min profiles): [Quiz - Production](https://github.com/FlorianBruniaux/claude-code-ultimate-guide/tree/main/quiz#advanced-categories) - Categories: MCP, production, advanced, learning, ecosystem
+     - Security-focused: [Quiz - Security Hardening](https://github.com/FlorianBruniaux/claude-code-ultimate-guide-landing/tree/main/questions/13-security-hardening) - Attack techniques, CVEs, campaigns, scanning tools
   - Cheat sheet: [Printable cheatsheet](https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/cheatsheet.md)
-   - Full guide: [Ultimate Guide (11K+ lines)](https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/ultimate-guide.md)
+   - Full guide: [Ultimate Guide (22K+ lines)](https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/ultimate-guide.md)

 4. **Section-specific links**: When referencing specific sections, use GitHub line anchors:
   - Format: `https://github.com/FlorianBruniaux/claude-code-ultimate-guide/blob/main/guide/ultimate-guide.md#L{line_number}`
@ -301,14 +307,14 @@ Begin by asking about preferred language.
 2. Simplify adaptive logic: Use static profiles from onboarding_matrix (ignore `adaptive` section, use `core` topics only)
 3. Manually paste reference.yaml content if WebFetch fails (or use fallback roadmap)

-**Localization status (v3.23.0):**
+**Localization status (v3.32.2):**
 - Core guide content: **English only**
- v3.21-3.22 topics: **English only** (dual_instance, git_mcp, sandbox_native, config_hierarchy, mcp_secrets)
+- v3.21-3.32 topics: **English only** (dual_instance, git_mcp, sandbox_native, config_hierarchy, mcp_secrets, plan_pipeline)
 - French/Spanish onboarding: Claude translates on-the-fly from English sections
 - **Limitation**: Translations not verified by native speakers, may have inaccuracies or awkward phrasing
- Quiz: English only (274 questions)
+- Quiz: English only (271 questions)

-**If translation quality is critical**: Recommend English onboarding for best accuracy, especially for technical v3.21-3.22 content.
+**If translation quality is critical**: Recommend English onboarding for best accuracy, especially for technical v3.21-3.32 content.
 ```

 ---