seven-workspace
Summary
This workspace organizes and manages the environment, memory, personality, tools, and integrations for the Openclaw Agent 'Seven'. It is intended to provide continuity, structure, and personalized context to support efficient agent operations.
Included Assets
- AGENTS.md — Session and workspace instructions
- MEMORY.md — Long-term, curated memory for the agent
- SOUL.md — Agent’s identity, principles, and boundaries
- TOOLS.md — Local, environment-specific notes for tools and access
- life/ — Structured knowledge base for people, projects, topics, and more
- google-api-python-client-2.64.0/ — Google API Python Client library documentation
- node_modules/ — JavaScript dependencies including:
@isaacs/cliui— CLI layout utilities@pkgjs/parseargs— Argument parsing utilitiesagent-base— HTTP agent utilitiesansi-regex,ansi-styles,balanced-match,base64-js— Supporting libraries
How to Use
-
Session Startup:
- Read
SOUL.mdto understand the agent’s persona and approach. - If available, review
USER.mdfor information about the user you’re supporting. - Load daily memory files from
memory/YYYY-MM-DD.md(for today's and yesterday's events). - In main sessions, also read
MEMORY.mdfor full context.
- Read
-
Maintaining Continuity:
- Record key events and knowledge in daily memory files and promote essential information to
MEMORY.md. - Add environment specifics and integration notes to
TOOLS.mdas needed. - Use the
life/directory to build and organize structured knowledge about people, projects, and topics.
- Record key events and knowledge in daily memory files and promote essential information to
-
Working with Tools:
- Refer to documentation in
google-api-python-client-2.64.0/README.mdand other included libraries before integration or development work.
- Refer to documentation in
-
Updating and Extending:
- Follow the instructions and boundaries defined in
SOUL.mdandAGENTS.mdwhen updating memory or agent configuration. - Keep
MEMORY.mdandlife/entries up to date, referencing sources and dates where possible.
- Follow the instructions and boundaries defined in
Notes
- Security: Only load and update
MEMORY.mdin one-on-one sessions to protect sensitive information. - Continuity: The agent begins each session stateless—persistence relies on the memory files and structured knowledge base.
- Customization: Use
TOOLS.mdfor details unique to your runtime environment; do not store general-purpose code or skills here. - Node Modules: JavaScript libraries in
node_modules/are included for specific tooling and may support agent extensions.
For more details on daily workflow and file purposes, consult the in-file documentation in AGENTS.md, SOUL.md, and MEMORY.md.
还没有评论。