openclaw-memory

Summary

This workspace provides a structured memory and workflow system for OpenClaw, tailored for advanced project management in animation or media production pipelines. It includes detailed memory protocols, asset management scripts, standardized knowledge on agents, tools, and step-by-step skill integrations for shot mapping, config building, and AI generation.

Included Assets

• Documentation
  • AGENTS.md: Workspace/agent behavior, memory, and operational rules
  • MEMORY.md: Long-term distilled workspace and user memory
  • SOUL.md: Agent principles, values, and personality definition
  • TOOLS.md: Local project and command references
• Memory System
  • memory/ARCHIVE/YYYY-MM-DD.md: Archived daily memory logs
  • memory/heartbeat-state.json: Memory maintenance and system health status
• Skills (skills/)
  • asset-library-manager/: Asset filtering and semantic search (docs, shell scripts)
  • pdf-storyboard-parser/: PDF storyboard parsing to images and text (docs, Python scripts, install scripts)
  • ai-shot-generator/: ZenStudio-based shot image/video AI generation (docs)
  • shot-config-builder/: Integrates assets, mapping, and storyboards into shot configs (docs)
  • shot-mapping-parser/: Excel shot breakdown to JSON mapping (docs)
  • Each skill provides a SKILL.md with workflow and integration details
• Scripts
  • Asset scanning and search: Bash scripts in skills/asset-library-manager/scripts/
  • PDF storyboard parsing: Python and shell scripts in skills/pdf-storyboard-parser/scripts/

How to Use

1. Boot Up & Get Context

   • Follow steps and etiquette described in AGENTS.md. On initial run, read SOUL.md, USER.md (if present), and latest daily memories.
   • Use MEMORY.md for ongoing distilled knowledge and workflow context.

2. Memory Management

   • Capture daily work/process notes in memory/YYYY-MM-DD.md.
   • Regularly promote key learnings or preferences into MEMORY.md. Prune memory: archive after 7 days, delete after 30.

3. Asset Management

   • Use skills/asset-library-manager/scripts/scan_assets.sh to scan, filter, and validate project assets.
   • Use skills/asset-library-manager/scripts/search_asset.sh to search assets with exact, loose, or fuzzy matching.

4. PDF Storyboard Extraction

   • Install requirements with the provided install script in skills/pdf-storyboard-parser/scripts/.
   • Use parse_pdf_storyboard.py or related tools to extract frames and descriptions from PDF storyboards for animation projects.

5. Skill Workflows

   • See each skill’s SKILL.md for input/output definitions and data integration steps—for example, parsing shot mappings, building config files, or using AI for content generation.
   • Follow the specified standard folder structures and conventions for each project.

6. Heartbeat & Maintenance

   • System health is tracked in memory/heartbeat-state.json.
   • Regularly review, summarize, and maintain the memory system as outlined.

Notes

• All workspace behaviors, memory pruning, and asset filtering rules are clearly documented in respective .md and script files.
• Workflows rely on Python (≥3.7), Bash, and standard data processing tools (pdf2image, pytesseract, pandas, etc.).
• Asset semantic analysis and search rely partly on external “image” tools and require manual tagging/validation during setup.
• Ensure workspace folder structures and naming conventions match sample project paths and filters in the documentation for best results.
• Heartbeat and system checks (output, disk, git) are automated; check heartbeat-state.json for status.
• Skills are modular—refer to each SKILL.md for clear scope, dependencies, and integration.
• Sensitive or external operations (remote publishing, destructive actions) require explicit user confirmation per AGENTS.md and SOUL.md.
• For non-Chinese speakers: operational rules and documentation are primarily in Chinese.

This workspace provides a complete, maintainable memory and workflow system for complex creative or media production pipelines in OpenClaw.