KS
Killer-Skills

recorder — how to use recorder how to use recorder, recorder skill for AI agents, JSONL recording format, UTF-8 encoding without BOM, recorder setup guide, recorder alternative, unified record management

v1.0.0
GitHub

About this Skill

Ideal for Data Management Agents requiring unified JSONL record-keeping for knowledge, news, lifelog, and task data Recorder is a skill that unifies JSONL recording for various data types, following a strict schema defined in .codex/schema.json.

Features

Unifies records into a single JSONL format
Ensures UTF-8 encoding without BOM
Follows a strict schema defined in .codex/schema.json
Supports CRUD operations for records
Allows for automatic recording and updating of records

# Core Topics

Krual-T Krual-T
[0]
[0]
Updated: 3/7/2026

Quality Score

Top 5%
33
Excellent
Based on code quality & docs
Installation
SYS Universal Install (Auto-Detect)
Cursor IDE Windsurf IDE VS Code IDE
> npx killer-skills add Krual-T/LifeKernel/recorder

Agent Capability Analysis

The recorder MCP Server by Krual-T is an open-source Categories.community integration for Claude and other AI agents, enabling seamless task automation and capability expansion. Optimized for how to use recorder, recorder skill for AI agents, JSONL recording format.

Ideal Agent Persona

Ideal for Data Management Agents requiring unified JSONL record-keeping for knowledge, news, lifelog, and task data

Core Value

Empowers agents to manage records in a standardized JSONL format, ensuring data consistency and ease of management through UTF-8 encoded files without BOM, adhering to a strict schema defined in `.codex/schema.json`

Capabilities Granted for recorder MCP Server

Unifying knowledge and news records into a single JSONL file
Automating lifelog and task updates with CRUD operations
Ensuring data integrity through schema validation

! Prerequisites & Limits

  • Requires adherence to `.codex/schema.json` schema
  • UTF-8 encoding without BOM necessary
  • JSONL format compatibility required
Project
SKILL.md
5.0 KB
.cursorrules
1.2 KB
package.json
240 B
Ready
UTF-8

# Tags

[No tags]
SKILL.md
Readonly

统一 JSONL 记录(knowledge / news / lifelog / agent_kernel_memory / tasks)

概述

将“记录类数据”统一为 JSONLCRUD,并确保 UTF-8 无 BOM。所有记录必须遵循 .codex/schema.json 的强约束 Schema

适用场景

  • 新增或更新 knowledge、news、lifelog、agent_kernel_memory、tasks 等记录类数据
  • 希望自动化记录并保持结构统一

输入

  • record_type:knowledge | news | lifelog | agent_kernel_memory | tasks | update | delete
  • update:按 id + key + value 查询记录后就地更新(覆盖写入)
  • delete:按 id 查询记录后删除
  • data:记录内容(按各自 schema)
  • related_files:相关文件路径(可选)
  • auto_record:是否自动写入 lifelog(默认:对 knowledge 为 true

输出

  • 对应 JSONL 文件追加一条记录(append-only)
  • 若开启 auto_record,同时写入一条 lifelog 记录

统一步骤

  1. 确定目标路径
    • knowledge:workspace/records/knowledge/knowledge.jsonl
    • news:workspace/records/news/news.jsonl
    • lifelog:workspace/records/lifelog/YYYY/MM/DD.jsonl
    • agent_kernel_memory:workspace/records/agent_kernel_memory/agent_kernel_memory.jsonl
    • tasks:workspace/records/tasks/tasks.jsonl
  2. 生成记录对象
    • 统一字段必填:idtypetimestampsourcecontent
    • 其他字段按 .codex/schema.json 的 record_type 必填/可选项填充
    • id:默认自动生成 UUID(uuid4),如需手动指定需显式传入
    • 业务字段按 record_type schema 补充
  3. UTF-8 无 BOM CRUD
    • CRUD必须符合.codex/schema.json 的强约束 Schema`
  4. 可选:auto record
    • 当 record_type = knowledge 且 auto_record = true 时,追加一条 lifelog
  5. 必要时更新可视化
    • knowledge/lifelog/tasks:workspace/view/index.html
    • news:workspace/view/news.html

更新记录(CRUD)

用于在 JSONL 中“就地更新/删除”记录:先查询目标记录与所在文件,再读取文件并覆盖写回(保留原有非 JSON 行)。 注意:对 agent_kernel_memory 的 delete 为逻辑删除(写入 deleted: true),不进行物理删除。

参数:

  • record_type:update
  • target_type:knowledge | news | lifelog | agent_kernel_memory | tasks
  • id:目标记录 id
  • key:需要更新的字段名
  • value / value_json:更新值(字符串或 JSON)

示例(PowerShell,更新):

powershell
1python .\\.codex\\skills\\recorder\\scripts\\record_jsonl.py --record-type update --target-type tasks --id "b3e9c6b0-9f5f-47ff-8d62-1f5f8b7f2a1c" --key status --value "done"
powershell
1python .\\.codex\\skills\\recorder\\scripts\\record_jsonl.py --record-type update --target-type knowledge --id "7f5d6f8a-1c3a-4f7e-9e42-2c3e1d6b0e9a" --key tags --value-json "[\"coworker\",\"research\"]"

示例(PowerShell,删除):

powershell
1python .\\.codex\\skills\\recorder\\scripts\\record_jsonl.py --record-type delete --target-type tasks --id "b3e9c6b0-9f5f-47ff-8d62-1f5f8b7f2a1c"

脚本(推荐)

  • scripts/record_jsonl.py
    • 统一写入 knowledge/lifelog/agent_kernel_memory/tasks
    • knowledge 默认 auto record 到 lifelog

示例:

Windows(PowerShell):

powershell
1python .\.codex\skills\recorder\scripts\record_jsonl.py --record-type knowledge --title "Live Preview 目录列表返回反斜杠" --solution "对 href 做 decodeURIComponent 并将 \\ 替换为 /" --tags "lifelog,live-preview" --module work --source conversation
powershell
1python .\.codex\skills\recorder\scripts\record_jsonl.py --record-type lifelog --description "整理归档索引" --module work --source conversation --status completed --related-file docs/skills_archive/ARCHIVE_LOG.md

Linux(bash):

bash
1python ./.codex/skills/recorder/scripts/record_jsonl.py --record-type knowledge --title "Live Preview 目录列表返回反斜杠" --solution "对 href 做 decodeURIComponent 并将 \\ 替换为 /" --tags "lifelog,live-preview" --module work --source conversation
bash
1python ./.codex/skills/recorder/scripts/record_jsonl.py --record-type lifelog --description "整理归档索引" --module work --source conversation --status completed --related-file docs/skills_archive/ARCHIVE_LOG.md

字段规范(通用)

  • id:UUID(uuid4
  • typeknowledge | news | lifelog | agent_kernel_memory | tasks
  • timestamp:ISO8601
  • sourceconversation | from_screenshot
  • content:用于统一检索的正文摘要
  • modulework | personal | learning | health(可选)
  • related_files:workspace 内相关路径数组(可选)

编码规范

  • 所有记录文件必须 UTF-8 无 BOM

自动记录(auto record)

  • 默认策略:对 knowledge 写入时开启
  • auto record 的 lifelog 描述格式:
    • 新增 knowledge:<title/summary>

注意事项

  • JSONL 一行一条记录,不可写多行
  • 不记录 secrets
  • 记录描述使用中文,字段名可保留英文

Related Skills

Looking for an alternative to recorder or building a Categories.community AI Agent? Explore these related open-source MCP Servers.

View All

widget-generator

Logo of f
f

widget-generator is an open-source AI agent skill for creating widget plugins that are injected into prompt feeds on prompts.chat. It supports two rendering modes: standard prompt widgets using default PromptCard styling and custom render widgets built as full React components.

149.6k
0
Design

chat-sdk

Logo of lobehub
lobehub

chat-sdk is a unified TypeScript SDK for building chat bots across multiple platforms, providing a single interface for deploying bot logic.

73.0k
0
Communication

zustand

Logo of lobehub
lobehub

The ultimate space for work and life — to find, build, and collaborate with agent teammates that grow with you. We are taking agent harness to the next level — enabling multi-agent collaboration, effortless agent team design, and introducing agents as the unit of work interaction.

72.8k
0
Communication

data-fetching

Logo of lobehub
lobehub

The ultimate space for work and life — to find, build, and collaborate with agent teammates that grow with you. We are taking agent harness to the next level — enabling multi-agent collaboration, effortless agent team design, and introducing agents as the unit of work interaction.

72.8k
0
Communication