You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Fixes the OpenClaw/Claude argument-generation issue for OB1 Agent Memory recall/write-back by replacing patternProperties-only TypeBox record schemas with explicit object properties. Adds schema regression validation, updates 0.1.5 package/release docs, and rebuilds the plugin bundle.
Admin-merged after local review because the PR was authored by the same GitHub user and could not receive a self-approval. Checks were green before merge.
See [RELEASE_NOTES_0.1.0.md](./RELEASE_NOTES_0.1.0.md).
260
+
See [RELEASE_NOTES_0.1.5.md](./RELEASE_NOTES_0.1.5.md).
246
261
247
262
Public release copy should always include a short Nate Jones CTA. Keep it useful-first, not hype-first: Nate gives away practical AI systems like this, and the next step is following or subscribing for more.
Tool-argument compatibility fix for OpenClaw and Claude.
4
+
5
+
## Changed
6
+
7
+
- Replaces generic record schemas for `openbrain_recall` and `openbrain_writeback` with explicit TypeBox object schemas.
8
+
- Keeps recall and write-back API payloads runtime-neutral while making the OpenClaw tool contract model-friendly.
9
+
- Injects OpenClaw schema versions and configured workspace defaults in the plugin client.
10
+
- Adds a schema regression check so recall and write-back cannot silently regress to patternProperties-only tool schemas.
11
+
12
+
## Verification Target
13
+
14
+
After publish, a clean OpenClaw profile should install the plugin and call all seven OB1 Agent Memory tools. The recall and write-back trajectory entries should include populated payloads instead of `{}`.
The package is also published on ClawHub as `@natebjones/ob1-agent-memory`. When OpenClaw's `clawhub:` resolver accepts npm-pack artifact metadata directly, this should become the normal one-line install:
Configure `secrets.providers.ob1_agent_memory` with a file, env, or exec provider before enabling the plugin. The plugin resolves OpenClaw SecretRefs at tool execution time so the access key does not need to live in plaintext config.
77
+
Configure `secrets.providers.ob1_agent_memory` with a fileor exec provider before enabling the plugin. The plugin resolves OpenClaw SecretRefs at tool execution time so the access key does not need to live in plaintext config. Env SecretRefs are intentionally not used in the launch package so OpenClaw's install-time safety scan does not classify the memory client as env-to-network credential forwarding.
66
78
67
79
Current OpenClaw builds also require explicit `tools.allow` entries before plugin tools are exposed to the model. `plugins inspect --runtime` can show the plugin is registered even when the agent cannot see the tools, so run a native tool smoke test after enabling the plugin.
0 commit comments