Skill Manager

Task Objectives

• This Skill is used to: Provide unified configuration and log storage capabilities for other skills
• Capabilities include:
  1. Store skill configuration information
  2. Store skill operation logs
  3. Query configurations and logs by skill name
  4. Manage stored data of all skills
• Trigger condition: When actively referenced by other skills for data storage or retrieval

Usage Scenarios

Scenario 1: Skill stores its own configuration

from skill_manager import SkillStorage

# 创建存储实例
storage = SkillStorage(data_path="/workspace/projects/skill-data.json")

# 存储配置
config = {
    "deploy_mode": "simple",
    "output_path": "/path/to/output.log",
    "timestamp": "2024-01-22 12:00:00"
}
storage.save_config("my-skill", config)

Scenario 2: Skill stores operation logs

from skill_manager import SkillStorage

storage = SkillStorage(data_path="/workspace/projects/skill-data.json")

logs = [
    {"time": "2024-01-22 12:00:00", "level": "INFO", "message": "开始执行"},
    {"time": "2024-01-22 12:05:00", "level": "INFO", "message": "执行完成"}
]
storage.save_logs("my-skill", logs)

Scenario 3: Skill reads data from other skills

from skill_manager import SkillStorage

storage = SkillStorage(data_path="/workspace/projects/skill-data.json")

# 读取其他技能的配置
other_config = storage.get_config("other-skill")

# 读取其他技能的日志
other_logs = storage.get_logs("other-skill")

Core Function Description

Functions Processable by Agents

• API usage consultation: Explain how to use each function of the Skill Manager
• Data format suggestions: Recommend appropriate data structures based on storage requirements
• Data analysis: Analyze stored configurations and logs to identify patterns or issues
• Storage management: View all stored skills and clean up expired data

Functions Implemented by Scripts

• SkillStorage Class: Complete storage and retrieval API, see scripts/skill_manager.py

  • save_config(skill_name, config)

    : Store skill configuration

  • save_logs(skill_name, logs)

    : Store skill logs

  • save(skill_name, config, logs)

    : Store configuration and logs simultaneously

  • get_config(skill_name)

    : Retrieve skill configuration

  • get_logs(skill_name)

    : Retrieve skill logs

  • get_all()

    : Retrieve all skill data

  • list_skills()

    : List all stored skills

  • delete(skill_name)

    : Delete skill data

Data Storage Format

{
  "skill-name-1": {
    "config": {
      "key1": "value1",
      "key2": "value2"
    },
    "logs": [
      {"time": "2024-01-22 12:00:00", "message": "日志1"},
      {"time": "2024-01-22 12:05:00", "message": "日志2"}
    ],
    "last_updated": "2024-01-22 12:05:00"
  },
  "skill-name-2": {
    "config": {},
    "logs": [],
    "last_updated": "2024-01-22 12:10:00"
  }
}

Resource Index

• Core Module: See scripts/skill_manager.py (Complete implementation of the SkillStorage class)
• API Specification: See references/api_spec.md (Detailed API documentation and usage examples)

Notes

• The Skill Manager is a service-type Skill that does not run actively and is called by other skills
• It is recommended to use

  /workspace/projects/skill-data.json

  as the storage path for unified management
• The data formats of configurations and logs are defined by the calling skills, and the Skill Manager only handles storage
• Each storage operation updates the

  last_updated

  timestamp
• Supports incremental updates and will not overwrite existing configurations or logs (unless explicitly specified to overwrite)

Best Practices

• Use consistent naming conventions for skill names (e.g., lowercase letters with hyphens)
• It is recommended to include necessary metadata in configuration data (such as creation time, version number, etc.)
• It is recommended to include timestamps and log levels in log data
• Regularly clean up unused skill data to prevent excessive storage file size
• Read configurations when the skill starts and record logs during execution
• Agents can assist in querying and analyzing stored data to identify skill collaboration opportunities

Usage Examples

Example 1: Skill integrates Skill Manager

import sys
sys.path.insert(0, '/workspace/projects/skill-manager/scripts')
from skill_manager import SkillStorage

# 初始化存储
storage = SkillStorage(data_path="/workspace/projects/skill-data.json")

# 存储技能配置
skill_config = {
    "mode": "production",
    "output_dir": "/workspace/output",
    "retry_count": 3
}
storage.save_config("my-awesome-skill", skill_config)

# 记录运行日志
execution_logs = [
    {"time": "2024-01-22 10:00:00", "level": "INFO", "message": "开始执行"},
    {"time": "2024-01-22 10:00:05", "level": "INFO", "message": "加载配置"},
    {"time": "2024-01-22 10:00:10", "level": "INFO", "message": "执行完成"}
]
storage.save_logs("my-awesome-skill", execution_logs)

Example 2: Query skill collaboration data

from skill_manager import SkillStorage

storage = SkillStorage(data_path="/workspace/projects/skill-data.json")

# 查看所有已存储的技能
all_skills = storage.list_skills()
print(f"已存储的技能: {all_skills}")

# 读取特定技能的配置
config = storage.get_config("dev-observability")
print(f"配置: {config}")

# 读取特定技能的日志
logs = storage.get_logs("dev-observability")
print(f"最近日志: {logs[-5:]}")

Example 3: Agent analyzes stored data

1. Read all skill data
2. Analyze configuration associations between skills
3. Identify shared dependencies or data
4. Provide optimization suggestions