agent-harness-construction

原文涵盖的主题

• Agent Harness Construction
• Core Model
• Action Space Design
• Granularity Rules
• Observation Design
• Error Recovery Contract
• Context Budgeting
• Architecture Pattern Guidance

Use this skill when you are improving how an agent plans, calls tools, recovers from errors, and converges on completion. Agent output quality is constrained by: 1. Action space quality 2. Observation quality 3. Recovery quality 4. Context budget quality 1. Use stable, explicit tool names. 2. Keep inputs schema-first and narrow. 3. Return deterministic output shapes. 4. Avoid catch-all tools unless isolation is impossible. Use micro-tools for high-risk operations (deploy, migration, permissions). Use medium tools for common edit/read/search loops. Use macro-tools only when round-trip overhead is the dominant cost. Every tool response should include: status: success|warning|error summary: one-line result nextactions: actionable follow-ups artifacts: file paths / IDs For every error path, include: root cause hint safe retry instruction explicit stop condition 1. Keep system prompt minimal and invariant. 2. Move large guidance into skills loaded on demand. 3. Prefer references to files over inlining long documents. 4. Compact at phase boundaries, not arbitrary token thresholds. ReAct: best for exploratory tasks with uncertain path. Function-calling: best for structured deterministic f

这里只展示节选，请以完整原文和当前要求为准。