القائمة
Enterprise context and session management with token budget optimization and state persistence
يُحدَّد Prompt الذي يراجع المصدر أولًا بشكل افتراضي. يمكنك التبديل إلى أمر مباشر أو تنزيل نسخة محلية.
اقرأ SKILL.md وأي ملفات مرافقة يعرضها SkillsMP قبل أن تقرر التثبيت.
التثبيت باستخدام Codex أو Claude انسخ هذا Prompt والصقه في Codex أو Claude أو مساعد آخر ليراجع صفحة Skill ويثبّتها لك.
يتجاوز الأمر المباشر Prompt المخصّص للمراجعة. افحص المصدر قبل تشغيله.
npx skills add https://github.com/majiayu000/claude-skill-registry --skill moai-foundation-contextيبقى الأمر في سطر واحد. مرّر أفقيًا لمراجعته كاملًا قبل النسخ.
تفضّل نسخة محلية؟ نزّل الملفات المتاحة حاليًا لدى SkillsMP.
LLM token logprobs and calibration. Per-decision confidence, ECE, Brier, reliability diagrams, low-confidence triage.
Analyze LLM token logprobs and calibration. Use for per-decision confidence, ECE, Brier scores, reliability diagrams, and low-confidence triage.
回顾最近 N 天的 Claude Code 使用记录——扫描原始会话数据，按主题分组汇总"我都做了什么"，并从个人操作系统视角输出模式、风险与增删建议。当用户说 /recap、"看看我这几天做了什么"、"回顾一下我最近的会话"、"这两天我用 claude 干了啥"、"活动回顾" 时使用。
استنادا إلى تصنيف SOC المهني
عرض SKILL.md
| name | moai-foundation-context |
| aliases | ["moai-foundation-context"] |
| description | Enterprise context and session management with token budget optimization and state persistence |
| version | 3.0.0 |
| modularized | false |
| category | foundation |
| replaces | moai-core-context-budget, moai-core-session-state |
| deprecated_names | {"moai-foundation-context":{"deprecated_in":"v0.32.0","remove_in":"v0.35.0","message":"Use moai-foundation-context instead"}} |
| tags | ["foundation","enterprise","framework","context","session","budget","architecture"] |
| updated | 2025-11-27T00:00:00.000Z |
| status | active |
Unified context optimization and session state management for Claude Code with 200K token budget management, session persistence, and multi-agent handoff protocols.
Core Capabilities:
When to Use:
Key Principles (2025):
/clear every 1-3 messages for SPEC workflowsPattern 1: Progressive File Loading
# Load files by priority tiers
Tier 1: CLAUDE.md, config.json (always loaded)
Tier 2: Current SPEC and implementation files
Tier : Related modules dependencies
Tier : Reference documentation (on-demand)
Pattern 2: Context Checkpointing
/clear to reset sessionPattern 3: MCP Context Continuity
# Preserve MCP agent context across /clear
agent_id = mcp_agent.get_id()
# After /clear:
resume_mcp_agent(agent_id) # Restores full context
Concept: Strategic allocation and monitoring of 200K token context window.
Allocation Strategy:
200K Token Budget Breakdown:
├── System Prompt & Instructions: ~15K tokens (7.5%)
│ ├── CLAUDE.md: ~8K
│ ├── Command definitions: ~4K
│ └── Skill metadata: ~3K
├── Active Conversation: ~80K tokens (40%)
│ ├── Recent messages: ~50K
│ ├── Context cache: ~20K
│ └── Active references: ~10K
├── Reference Context (Progressive Disclosure): ~50K (25%)
│ ├── Project structure: ~15K
│ ├── Related Skills: ~20K
│ └── Tool definitions: ~15K
└── Reserve (Emergency Recovery): ~55K tokens (27.5%)
├── Session state snapshot: ~10K
├── TAGs and cross-references: ~15K
├── Error recovery context: ~20K
└── Free buffer: ~10K
Monitoring Thresholds:
def monitor_token_budget(context_usage: int):
"""Real-time token budget monitoring with automatic actions."""
usage_percent = (context_usage / 200000) * 100
if usage_percent > 85:
# Critical: Emergency compression
trigger_emergency_compression()
execute_clear_command()
elif usage_percent > 75:
# Warning: Start progressive disclosure
defer_non_critical_context()
warn_user_approaching_limit()
elif usage_percent > 60:
# Monitor: Track growth patterns
track_context_growth()
Use Case: Prevent context overflow in long-running SPEC-First workflows.
/clear StrategyConcept: Proactive context clearing at strategic checkpoints to maintain efficiency.
Clear Execution Rules:
MANDATORY /clear Points:
├── After /moai:1-plan completion (saves 45-50K tokens)
├── Context > 150K tokens (prevents overflow)
├── Conversation > 50 messages (removes stale history)
├── Before major phase transitions (clean slate)
└── Model switches (Haiku ↔ Sonnet handoffs)
Implementation:
def should_execute_clear(context: dict) -> bool:
"""Determine if /clear should be executed."""
triggers = {
"post_spec_creation": context.get("spec_created", False),
"token_threshold": context.get("token_usage", 0) > 150000,
"message_count": context.get("message_count", 0) > 50,
"phase_transition": context.get("phase_changed", False)
}
return any(triggers.values())
Use Case: Maximize token efficiency across SPEC-Run-Sync cycles.
Concept: Maintain session continuity across interruptions with state snapshots.
Session State Architecture:
Session State Layers:
├── L1: Context-Aware Layer (Claude 4.5+ feature)
│ ├── Token budget tracking
│ ├── Context window position
│ ├── Auto-summarization triggers
│ └── Model-specific optimizations
├── L2: Active Context (current task, variables, scope)
├── L3: Session History (recent actions, decisions)
├── L4: Project State (SPEC progress, milestones)
├── L5: User Context (preferences, language, expertise)
└── L6: System State (tools, permissions, environment)
State Snapshot Structure:
{
"session_id": "sess_uuid_v4",
"model": "claude-sonnet-4-5-20250929",
"created_at": "2025-11-24T10:30:00Z",
"context_window": {
"total": 200000,
"used": 85000,
"available": 115000,
"position_percent": 42.5
},
"persistence": {
"auto_load_history": true,
"context_preservation": "critical_only",
"cache_enabled": true
},
"work_state": {
"current_spec": "SPEC-001",
"phase": "implementation",
"completed_steps": ["spec_complete", "architecture_defined"]
}
}
Use Case: Resume long-running tasks after interruptions without context loss.
Concept: Seamless context transfer between agents with minimal token overhead.
Handoff Package:
{
"handoff_id": "uuid-v4",
"from_agent": "spec-builder",
"to_agent": "tdd-implementer",
"session_context": {
"session_id": "sess_uuid",
"model": "claude-sonnet-4-5-20250929",
"context_position": 42.5,
"available_tokens": 115000,
"user_language": "ko"
},
"task_context": {
"spec_id": "SPEC-001",
"current_phase": "implementation",
"completed_steps": ["spec_complete", "architecture_defined"],
"next_step": "write_tests"
},
"recovery_info": {
"last_checkpoint": "2025-11-24T10:25:00Z",
"recovery_tokens_reserved": 55000,
"session_fork_available": true
}
}
Handoff Validation:
def validate_handoff(handoff_package: dict) -> bool:
"""Validate handoff package integrity."""
# Token budget check
available = handoff_package['session_context']['available_tokens']
if available < 30000: # Minimum safe buffer
trigger_context_compression()
# Agent compatibility check
if not can_agents_cooperate(
handoff_package['from_agent'],
handoff_package['to_agent']
):
raise AgentCompatibilityError("Agents cannot cooperate")
return True
Use Case: Efficient Plan → Run → Sync workflow execution.
Concept: Load context progressively based on relevance and need.
Progressive Summarization:
def progressive_summarization(context: str, target_ratio: float = 0.3):
"""Compress context while preserving key information."""
# Step 1: Extract key sentences (50K → 15K)
summary = extract_key_sentences(context, ratio=target_ratio)
# Step 2: Add pointers to original content
summary_with_refs = add_content_pointers(summary, context)
# Step 3: Store original for recovery
store_original_context(context, "session_archive")
return summary_with_refs # 35K tokens saved
Context Tagging:
# ❌ Bad (high token cost):
"The user configuration from the previous 20 messages..."
# ✅ Good (efficient reference):
"Refer to @CONFIG-001 for user preferences"
Use Case: Maintain context continuity while minimizing token overhead.
For detailed patterns and implementation strategies:
/clear immediately after SPEC creationmoai-cc-memory - Memory management and context persistencemoai-cc-configuration - Session configuration and preferencesmoai-core-workflow - Workflow state persistence and recoverymoai-cc-agents - Agent state management across sessionsmoai-foundation-trust - Quality gate integrationSession Initialization:
1. Initialize token budget (Pattern 1)
2. Load session state (Pattern 3)
3. Setup progressive disclosure (Pattern 5)
4. Configure handoff protocols (Pattern 4)
SPEC-First Workflow:
1. /moai:1-plan execution
↓
2. /clear (mandatory - saves 45-50K tokens)
↓
3. /moai:2-run SPEC-XXX
↓
4. Multi-agent handoffs (Pattern 4)
↓
5. /moai:3-sync SPEC-XXX
↓
6. Session state persistence (Pattern 3)
Context Monitoring:
Continuous:
├─ Track token usage (Pattern 1)
├─ Apply progressive disclosure (Pattern 5)
├─ Execute /clear at thresholds (Pattern 2)
└─ Validate handoffs (Pattern 4)
Status: Production Ready (Enterprise) Modular Architecture: SKILL.md + 6 modules Integration: Plan-Run-Sync workflow optimized Generated with: MoAI-ADK Skill Factory