← 返回 Skills 市场
124
总下载
0
收藏
1
当前安装
7
版本数
在 OpenClaw 中安装
/install opc-journal
功能描述
OPC200 Journal - A CLI-style single skill for One Person Company growth tracking. Record entries, analyze patterns from dreams/memory, detect milestones, and...
安全使用建议
This skill behaves like a local journal CLI and most visible code is consistent with that claim, but you should do two quick checks before installing or running it with real data: (1) Inspect the omitted utility files (utils/storage.py, utils/task_storage.py, maybe timezone) to confirm they do not perform any network calls, invoke unexpected binaries, or read unrelated system paths or environment variables. (2) Confirm that build_customer_dir and related storage functions properly sanitize customer_id and constrain files under a safe directory (e.g., ~/.openclaw/customers/...), to avoid path traversal or writing outside the intended area. Also be aware that analyze/insights return raw journal contents and signal counts to the calling LLM — this is expected for the skill but is a privacy risk if you don't trust the agent/LLM. If you cannot inspect the omitted files or do not want journal contents exposed to an LLM/agent, do not install or run this skill with real personal data.
功能分析
Type: OpenClaw Skill
Name: opc-journal
Version: 2.5.2
The opc-journal skill bundle is a well-structured local journaling and task-tracking tool. It implements several security best practices, including path sanitization to prevent traversal attacks (utils/storage.py), file locking for data integrity (utils/task_storage.py), and mandatory confirmation flags for destructive operations like deletion or archiving (scripts/commands/delete.py). The code contains no network-related imports, obfuscation, or unauthorized execution logic, and strictly adheres to its stated 'local-only' architecture.
能力评估
Purpose & Capability
Name/description (local CLI journal, analysis, milestones, insights) align with the included command modules: read/write/search/archive/analyze/insights/task persistence under a per-customer directory. No unrelated cloud credentials or binaries are requested.
Instruction Scope
SKILL.md and the visible command code focus on local file I/O and returning raw structured data to the caller (LLM). This is coherent, but the skill intentionally returns raw journal text and signal counts to the caller — a privacy-sensitive action (exposes potentially very sensitive personal data to the calling LLM). Also, some commands (insights/analyze) include keyword regexes (e.g., 'burnout', 'overwhelm') which surface mental-health-related signals; the code claims not to draw conclusions, but it does surface sensitive indicators for an LLM to interpret.
Install Mechanism
No install spec; instruction-only/CLI files are present and nothing will be downloaded or injected during install. This is the lowest install risk.
Credentials
No environment variables, credentials, or config paths are declared or required. However, several utility files referenced by the commands (utils/storage.py, utils/task_storage.py, utils/timezone.py) were omitted from the provided content — those could legitimately access environment variables, external paths, or networking. The lack of declared env vars is appropriate for the stated purpose, but missing utility files prevent a complete verification.
Persistence & Privilege
Skill is not always-enabled, does not request elevated system privileges, and appears to confine its I/O to a per-customer directory under the user's home. It does not modify other skills or global agent settings.
如何使用
- 确保已安装 OpenClaw(本地或 Docker 部署)
- 在对话框中输入安装命令:
/install opc-journal - 安装完成后,直接呼叫该 Skill 的名称或使用
/opc-journal触发 - 根据 Skill 的参数说明提供必要输入,即可获得结构化输出
版本历史
v2.5.2
- Expanded documentation for directory structure and command behaviors.
- Added detailed sections on security model, concurrency, and storage format.
- All destructive operations (`delete`, `archive --clear`) now require a `--force` flag for extra safety.
- Command outputs are now standardized to JSON with `status`, `result`, and `message` fields.
- Skill continues to extract only structural and keyword signals, delegating all interpretation to the caller.
v2.5.1
- Removed ARCHITECTURE_DECISION.md and AUDIT_REPORT_v2.5.0.md files.
- No changes to features, interface, or commands.
- No changes to local data storage or processing.
- No external dependencies added or removed.
v2.5.0
**Major update: New features for batch task management, entry deletion, and journal archiving.**
- Added `batch-task` command for creating multiple async tasks at once.
- Added `delete` command to remove journal entries by `entry_id`.
- Added `archive` command to archive all journal data, with option to clear after archiving.
- Improved modularity with new utility scripts for parsing, task storage, and timezone handling.
- Updated documentation to reflect new commands and clarify data handling philosophy.
v2.4.3
OPC Journal v2.4.3
Highlights
OPC Journal v2.4.3 finalizes the migration from the legacy multi-skill architecture into a single, streamlined CLI skill. It ships with persistent language localization, retroactive document translation, and full compatibility with OpenClaw v2026.4.11.
Changes
Single CLI Skill Architecture
Retired the opc-journal-suite coordinator + 5 sub-skills model.
Consolidated all capabilities into one skill (opc-journal) with 10 uniform sub-commands: init, record, search, export, analyze, milestones, insights, task, status, and update-meta.
Updated Dockerfile.light, CI workflows (skills-test.yml, publish-skills.yml), Makefile, test scripts, and deployment helpers to reference the new single skill.
Localization & Language Persistence
All user-facing output is now localized (Chinese by default, English available).
init auto-detects language from user input and persists the choice in journal_meta.json.
Every subsequent command reads the stored language setting and responds dynamically.
Dynamic Emotion Profiling
Replaced the static keyword-to-label mapping with a 4-dimension dynamic model (energy, tension, valence, fatigue).
Emotion output is now descriptive and contextual rather than rigidly categorized.
update-meta Command
Added update-meta --language <zh|en> --goals [...] --preferences {...}.
When the language changes, the command retroactively translates fixed template labels (charter headings, entry labels, metadata sections) across all existing memory/*.md files.
User-written content is never altered; only known boilerplate strings are swapped via precise regex rules.
OpenClaw v2026.4.11 Compatibility
Verified SKILL.md frontmatter compliance with the 2026.4.11 skill parser.
Confirmed scripts/main.py works correctly under command-dispatch: tool / tool: main / command-arg-mode: raw.
Installed and validated the skill in ~/.openclaw/skills/; openclaw skills check reports ready.
Removed 6 legacy suite skills from the local environment to eliminate shadowing.
Docs & Build Cleanup
Rewrote SKILL.md entirely in English, with no references to the old suite.
Eliminated all remaining opc-journal-suite references across docs/, .github/, scripts/, Makefile, and .coveragerc.
Added test_update_meta.py with zh→en and en→zh coverage.
Fixes
init arguments now correctly expose --language and --preferences.
record burnout keyword no longer prepends an unwanted leading space.
SKILL.md frontmatter format normalized to strict YAML (resolves parsing issues with some clawhub publish versions).
v2.4.2
OPC Journal v2.4.2 is a major architectural consolidation of the journaling experience for the OPC200 framework. We have retired the multi-skill opc-journal-suite model and rebuilt everything as a single, streamlined CLI skill under opc-journal.
This release delivers 10 unified sub-commands (init, record, search, export, analyze, milestones, insights, task, status, and the new update-meta), fully localized output with persistent language detection, and dynamic emotion profiling based on four behavioral dimensions. A key addition is update-meta --language, which lets existing users switch languages retroactively—automatically translating fixed document templates in their memory files while leaving their original content untouched.
v2.4.1
Version 2.4.1
- No changes detected from the previous version.
- Skill functionality, description, and documentation remain unchanged.
v2.4.0
**Summary:**
opc-journal v2.4.0 consolidates all OPC journal functionality into a single, streamlined CLI interface with enhanced local analysis features.
- Merged five previously separate skills into one unified CLI skill: opc-journal
- Centralized functionality accessible via subcommands (init, record, search, export, analyze, milestones, insights, task, status, help)
- Added new analyze and insights commands for pattern detection and OPC-specific interpretation based on dream and memory files
- Ensured all journal data is stored and processed locally, with no network calls
元数据
常见问题
OPC Journal 是什么?
OPC200 Journal - A CLI-style single skill for One Person Company growth tracking. Record entries, analyze patterns from dreams/memory, detect milestones, and... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 124 次。
如何安装 OPC Journal?
在 OpenClaw 或 Claude Code 对话框中运行命令「/install opc-journal」即可一键安装,无需额外配置。
OPC Journal 是免费的吗?
是的,OPC Journal 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。
OPC Journal 支持哪些平台?
OPC Journal 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。
谁开发了 OPC Journal?
由 coidea(@coidea)开发并维护,当前版本 v2.5.2。
推荐 Skills