bot-uichan/openclaw-workspace

Summary

A TypeScript-based OpenClaw workspace, offering practical agents, skills, and memory structure for automating workflows such as OCR-based discount reading, chat boundary enforcement, crypto demo-trading, Discord multi-mention, GitHub commenting, and Nostr relay search. Designed for hands-on integration and iterative automation, focusing on clarity and user privacy.

Included Assets

• Agents & Persona
  • AGENTS.md: Workspace operating guide, memory usage, and session routines.
  • SOUL.md: Defines the assistant's core philosophy, behaviors, and persona guidelines.
• Apps (under apps/)
  • discount-ocr: React + Tesseract.js PoC for OCR discount detection and calculation.
  • dlsite-play-cookie-helper: Playwright-based CLI helper to retrieve cookies for authentication flows.
  • dlsite-play-tui: A TypeScript TUI client for DLsite Play, with optional external cookie helper integration.
• Memory
  • Daily activity logs and state persistence (memory/YYYY-MM-DD.md, heartbeat-state.json) for operational continuity.
• Skills (under skills/)
  • bounded-agent-chat: Enforces boundaries in agent-to-agent or multi-session LLM conversations.
  • demo-crypto-trade: Guides heartbeat-based discretionary crypto paper trading with rationale and research requirements.
  • discord-mutual-mention: Safely mentions multiple users in Discord with proper ID formatting.
  • github-pr-issue-comment: Posts robustly formatted comments to GitHub Issues/PRs via CLI.
  • nostr-search: Guides data fetch protocols for Nostr relay connections.

How to Use

1. Workspace Bootstrapping
   • Start by reading AGENTS.md for workflow and session management procedures.
   • Review SOUL.md to understand the persona, trust boundaries, and operational style expected.
   • Check recent files in memory/ for contextual awareness and session state.
2. Project Apps
   • In apps/discount-ocr/:
     • Install dependencies and launch with npm run dev for local testing of the OCR PoC.
     • Review its README for detailed dev/build instructions and feature notes.
   • In apps/dlsite-play-cookie-helper/ and apps/dlsite-play-tui/:
     • Install, link, and follow the command documentation in each app's README to authenticate and run the DLsite TUI.
3. Implementing Skills
   • Refer to each skill’s SKILL.md for operational flow, command scripting, integration strategies, and rules (e.g., how to safely comment on GitHub, enforce chat boundaries, or notify on Discord).
   • Use provided scripts (such as GitHub commenting via shell) as templates or starting points for automation.
4. Add, Review, and Curate Memory
   • Log daily interactions and automation outcomes in memory/YYYY-MM-DD.md.
   • Curate important events, decisions, or context into long-term memory for reference during direct sessions.

Notes

• All workflow routines and memory persistence are file-based, supporting clear review and portability.
• No cloud sync, database, or external services are assumed except where mentioned for app/script compatibility (e.g., Tesseract.js for in-browser OCR, Playwright for automation, or gh for GitHub CLI).
• The workspace honors privacy, context separation, and operational clarity as defined in its persona and agent guides.
• Skills are designed for shell/CLI, local workflows, or API orchestration, not graphical/no-code manipulation.
• Japanese and English documentation may both be present, reflecting the primary use cases and intended audience.
• For sensitive commands (messaging, code review), both templates and cautionary notes are included to minimize operational accidents.

Use this workspace as a foundation for practical OpenClaw agent automation, dialogue management, live trading simulations, and communication tool integration, while following the clear behavioral and memory curation standards provided.