菜单
Context window budget management and selective capability loading for LLM token optimization.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
| name | context-budget |
| description | Context window budget management and selective capability loading for LLM token optimization. |
| version | 1.0.0 |
| triggers | ["context-warning","session-start","manual"] |
Purpose: Manage LLM context window as a finite resource. Ensure the right agents and skills are loaded for each task while preventing context overflow.
LLM context windows have hard token limits. Loading all 19 agents and 27 skills simultaneously consumes context that should be reserved for the actual work. This skill implements intelligent selective loading to maximize usable context for code and reasoning.
At the start of each task, analyze the user request to identify which domains are involved:
User: "Add JWT authentication to the API"
→ Detected domains: security, backend, testing
→ Load: security-reviewer, backend-specialist, tdd-guide
→ Load: security-practices, api-patterns, testing-patterns
Consult engine/loading-rules.json to determine which agents and skills to load:
rules.md, manifest.json, workflow-state.jsonsession-context.md, session-state.json| Resource | Default Limit | Rationale |
|---|---|---|
| Agents per session | 4 | Each agent ~200-500 tokens of context |
| Skills per session | 6 | Each skill ~200-800 tokens of context |
| Maximum loaded context | ~30% of window | Reserve 70% for code and reasoning |
When approaching context limits:
strategic-compact skill to summarize completed worksession-state.json before compaction| Task Type | Agents Loaded | Skills Loaded |
|---|---|---|
| Security audit | security-reviewer, architect | security-practices, architecture |
| Feature build | planner, architect, tdd-guide | plan-writing, testing-patterns, clean-code |
| Bug fix | build-error-resolver | debugging-strategies |
| Sprint planning | sprint-orchestrator, planner | plan-writing, brainstorming |
| Code review | code-reviewer, security-reviewer | clean-code, security-practices |
| Deployment | devops-engineer, reliability-engineer | deployment-procedures, docker-patterns |
engine/loading-rules.jsonreliability-engineer agentstrategic-compact skillPull request lifecycle domain knowledge — branch strategy detection, PR size classification, confidence-scored review, git-aware context, PR analytics, dependency management, and split/merge/describe operations.
Production readiness audit domains, weighted scoring criteria, and check specifications for the /preflight workflow.
Application scaffolding orchestrator. Creates full-stack applications from requirements, selects tech stack, coordinates agents.
Production deployment workflows, rollback strategies, and CI/CD best practices.
Internationalization and localization patterns for multi-language applications
Mobile UI/UX patterns for iOS and Android. Touch-first, platform-respectful design with React Native/Expo focus.