cache-monitor

To track static workspace file changes and protect your prompt cache, you can initialize SHA-256 hashing for core files and run checks to detect modifications. This skill compares content hashes and emits warnings to trigger re-evaluation when static-layer files change.

Your prompt cache breaks across multi-host environments when static-layer files like SOUL.md or IDENTITY.md are modified without detection. Hashing these assets and persisting state guards cache integrity by exposing unauthorized changes during session startup.

To initialize and verify static file hashes for workspace auditing, run the CLI with --init to establish baseline SHA-256 hashes, then execute --check to compare current states and generate a human-readable report of cache status.

Yes, you can use deterministic SHA-256 hash checks for static-layer workspace validation. This skill applies fixed checks on a defined set of static files and writes state to a JSON file, ensuring consistent auditing across sessions and hosts.

A limitation of monitoring prompt cache integrity with file hashing is that it only tracks a fixed set of predefined static files. Dynamic content or files outside the defined static-layer assets are not hashed, meaning unchecked modifications will not trigger cache warnings.

Yes, cache-monitor works with OpenClaw workspace startup scripts by automatically activating during agent bootstrap. It persists state to the OpenClaw workspace directory, ensuring prompt cache integrity is verified before multi-host sessions begin.