Friends of Clawd Discord is now live!
A place to hang out, share your Clawdis setups, ask questions, and help each other build cool stuff with AI assistants.
Join the crustacean crew:
Anthropic extended-thinking sessions recover better after prompt-cache expiry or Gateway restart, and richer MCP tool blocks get normalized before they poison history.
Matrix got less awkward: voice notes can be transcribed before mention gating, and native threads preserve reply/read context instead of flattening the conversation.
Plugin and ClawHub installs are sturdier: pinned GitHub skill commits, install-policy checks, and SQLite-backed install records for better update/repair behavior.
security.installPolicy lets operators approve or block skill/plugin installs with a trusted local command. It runs after staging and fails closed when enabled but unavailable.
Parallel is now bundled as a web_search provider. Use the zero-config free path or bring PARALLEL_API_KEY for the paid API, with dense excerpts built for LLM context.
Release naming changed: OpenClaw now uses YYYY.M.PATCH. That last number is a monthly patch counter, not the calendar day; old tags still compare correctly. This is to help support future critical fixes on prior releases.
my most useful codex skill is an automation to scan over past conversations and create/refine skills based on the session history. @openclaw now does this automatically with skill workshops 🦞🎓
Agents should learn repeated work, but not by silently rewriting future runs.
Skill Workshop turns reusable agent lessons into reviewable proposals you can tweak, apply, or reject before they become live skills. openclaw.ai/blog/openclaw-…
Agents should learn repeated work, but not by silently rewriting future runs.
Skill Workshop turns reusable agent lessons into reviewable proposals you can tweak, apply, or reject before they become live skills.