openclaw-interchange

What This Skill Does

The @openclaw/interchange skill serves as the foundational communication layer for the OpenClaw agent ecosystem. It standardizes how agents exchange information by using structured markdown files (.md) as a persistent, human-readable, and machine-verifiable transport layer. By implementing atomic file operations, it ensures that data integrity is maintained even during sudden agent failures or power interruptions. The library utilizes a combination of YAML frontmatter for metadata and raw markdown for content, allowing agents to store context, state, and instructions in a single, version-controlled format.

Installation

To integrate this core library into your environment, run the following command in your terminal or agent command prompt: clawhub install openclaw/skills/skills/frank-bot07/openclaw-interchange

Use Cases

• Agent State Synchronization: Maintain consistent agent state across restarts by writing progress to disk.
• Inter-Agent Communication: Pass structured tasks between different agents by writing to shared directories where multiple agents monitor for changes.
• Audit Logging: Create verifiable, hash-tracked logs of agent decisions that can be audited later.
• Configuration Management: Update skill configurations dynamically using YAML frontmatter without rewriting the entire logic file.

Example Prompts

1. "Check the current system status by reading the ops/status.md interchange file and report any critical errors found in the metadata."
2. "Write an updated task list for the research agent to the shared workspace using the interchange format, ensuring you include the schema-validated frontmatter."
3. "Initialize a new project state file using openclaw-interchange, setting the status to 'initialized' and including the current timestamp in the YAML metadata."

Tips & Limitations

• Atomicity: Always rely on the library's built-in write methods rather than standard file system write calls to ensure the tmp-fsync-rename process prevents file corruption.
• Locking: Utilize advisory locking when working in multi-agent environments to prevent race conditions when two agents attempt to write to the same interchange file simultaneously.
• Performance: While file-based interchange is highly reliable and persistent, it is intended for state synchronization and small data packets rather than high-throughput streaming. Keep files under 1MB for optimal performance. The library includes circuit breaker patterns; monitor these logs to detect if an agent is repeatedly failing to release file locks.