📦 Pattern Recognition — 模式识别
v1.0.0在操作、错误与资源中自动发现规律，生成模板并评估效率，给出优化建议。
0· 117·1 当前·1 累计
by 下载技能包
最后更新
2026/4/2
安全扫描
OpenClaw
安全
high confidenceThe skill's code and instructions are internally consistent with its stated purpose (learning from operations, errors, and resources) and do not request unrelated credentials or network access, though it will read and write files in the agent workspace which you should review before enabling.
评估建议
This skill appears to do what it says: it scans the agent workspace for operations, logs, and metrics and writes learned patterns and suggestions back to a patterns folder. Before installing, review the contents of /home/openclaw/.openclaw/workspace (memory, logs, metrics) to ensure no sensitive secrets or credentials would be processed or exposed. Note there is no network/exfiltration code in the included scripts, but the SKILL.md references a 'pattern' CLI that isn't provided — verify how the ...详细分析 ▾
ℹ 用途与能力
The name/description match the included Python scripts: both learner and analyzer operate on operation history, logs, and metrics. One minor inconsistency: SKILL.md shows CLI usage (pattern ...) but no install spec or CLI wrapper is provided in the package — the repo contains Python scripts but not an installed 'pattern' command. Functionality requested (reading workspace markdown/log/metrics) is aligned with the stated purpose.
ℹ 指令范围
Runtime instructions and the scripts read files under /home/openclaw/.openclaw/workspace (memory/*.md, logs/*.log, metrics/*.json) and write patterns/suggestions/logs back into a patterns directory. This is appropriate for pattern-learning but means the skill will process potentially sensitive agent memory and logs; there are no network calls or external endpoints in the code.
✓ 安装机制
No install spec (instruction-only) and only two Python scripts are included. This is low-risk from an install standpoint because nothing is downloaded or extracted from external URLs.
✓ 凭证需求
The skill requests no environment variables or external credentials. It hardcodes a workspace path (/home/openclaw/.openclaw/workspace) which is reasonable for an agent-local analysis tool but gives it access to whatever lives in that workspace.
✓ 持久化与权限
always is false and the skill does not modify other skills or global agent settings. It persists learned state and logs under the workspace patterns directory only.
安全有层次，运行前请审查代码。
运行时依赖
无特殊依赖
版本
latestv1.0.02026/4/2
Initial release of advanced pattern recognition and learning for OpenClaw. - New skill for identifying, learning, and applying patterns in operations, errors, and workflows - Includes pattern learning, template generation, efficiency analysis, and optimization suggestions - Provides CLI commands for learning from logs, generating templates, analyzing efficiency, and getting recommendations - Introduces modular components (Pattern Learner, Template Generator, Efficiency Analyzer) - Supports installation, configuration, and integration with memory, watchtower, and specific agents
● 无害
安装命令
点击复制官方npx clawhub@latest install pattern-recognition
镜像加速npx clawhub@latest install pattern-recognition --registry https://cn.longxiaskill.com