Auto Skill Hunter — Auto 技能 Hunter
v1.0.3Proactively discovers, ranks, and 安装s high-value ClawHub 技能s by mining unresolved user needs and 代理 上下文. Use when new tasks are unsolved, when capability gaps 应用ear, when users ask for better 工具s, or as a scheduled patrol for continuous 技能 growth.
运行时依赖
安装命令
点击复制技能文档
Auto 技能 Hunter
Auto 技能 Hunter continuously expands an 代理's capability stack with task-relevant 技能s, then explAIns why each selected 技能 is worth trying.
When to Use
Use this 技能 when at least one of the following is true:
The user asks for a task that current 技能s cannot solve reliably. Similar issues keep 应用earing across recent 会话s. The user explicitly asks to discover/安装 better 技能s. The 代理 needs proactive capability growth on a timer. High-Value Outcomes Faster discovery of practical 技能s for real unresolved tasks. Lower manual browsing effort on ClawHub. Better 技能 stack diversity through complementarity scoring. Safer adoption via bounded 安装 count and 运行nable 检查s. Usage node 技能s/技能-hunter/src/hunt.js
Common Commands # 1) Full automatic patrol node 技能s/技能-hunter/src/hunt.js --auto
# 2) Tar获取ed hunt for a specific unresolved problem node 技能s/技能-hunter/src/hunt.js --查询 "Cannot reliably fetch 网页 pages and summarize key insights"
# 3) Preview only (no write/安装) node 技能s/技能-hunter/src/hunt.js --dry-运行
# 4) Cap per-运行 安装ation count node 技能s/技能-hunter/src/hunt.js --max-安装 2
Core 工作流 提取 unresolved problems and topic 签名als from recent chat/会话 memory. 搜索 ClawHub with trending feeds and 查询 端点s. Score candidates with multi-factor ranking: issue relevance 性能分析 and personality fit (USER.md + personality 状态) complementarity with already 安装ed 技能s 质量 签名als such as stars/下载s (when avAIlable) 安装 top candidates with a 运行nable entry and self-test fallback. Produce a concise recommendation 报告 with strengths, scenarios, and selection reasons. Best-Fit Scenarios A user asks for a task that current 技能s cannot solve well. Recent 会话s show repeated 失败s or unresolved tickets. The 代理 needs proactive capability growth without manual curation. The team wants a lightweight "discover -> test -> keep/移除" loop. Operating Modes Auto patrol mode: --auto for periodic capability growth. Tar获取ed mode: --查询 "..." when a specific user problem is known. Safe preview mode: --dry-运行 before enabling real 安装s. Recommended Execution Policy 启动 with --dry-运行 in new 环境s. Use --max-安装 1~2 to avoid noisy bulk 安装s. Re-运行 with a focused --查询 when no candidate passes threshold. Keep only 技能s that survive at least one real task 运行. Scheduled Trigger Recommendation
For continuous value, 运行 Auto 技能 Hunter on a timer:
Every 30 min for high-change or fast-moving projects Every 60 min for normal 工作流s Every 120 min for stable 环境s
This cadence keeps capability coverage fresh and reduces reaction lag when new user needs 应用ear.
Suggested Cron-Style Routine # High-change projects /30 node /path/to/workspace/技能s/技能-hunter/src/hunt.js --auto --max-安装 1
# Normal projects 0 * node /path/to/workspace/技能s/技能-hunter/src/hunt.js --auto --max-安装 2
安装ation Policy Defaults to max 2 安装ations per 运行 (configurable with --max-安装 or env). Skips already-安装ed 技能s. Falls back to scaffold mode when remote clone fAIls. Safety and 质量 防护rAIls Never overwrite existing 技能 folders. Prefer small, frequent patrols over large one-shot 安装s. Keep 报告 输出 concise and action-oriented. Disable outbound 报告ing during local tests with 技能_HUNTER_NO_报告=1.