Setup & Installation
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
What This Skill Does
Maintains a running log of mistakes, wrong assumptions, and project-specific discoveries in a local learnings.md file. Entries are reviewed before complex tasks and promoted to permanent config when they become critical. Keeps session knowledge from resetting between conversations.
Unlike chat history or ad-hoc notes, entries are structured, dated, and surfaced automatically before tasks where they apply.
When to use it
- Recording that a project uses pnpm after a failed npm install
- Noting a Docker ARM64 workaround before it slips away
- Reviewing past mistakes before starting a complex database migration
- Logging a naming convention discovered mid-code-review
- Tracking a flaky test root cause so it isn't re-diagnosed next sprint
Example Workflow
Here's how your AI assistant might use this skill in practice.
User asks: Fix the failing CI build, it keeps breaking on Docker
- 1Read learnings.md to check for existing Docker-related entries
- 2Find a prior entry noting the ARM64 platform mismatch
- 3Apply --platform linux/amd64 to the build command
- 4Append a new entry capturing the specific CI variant of the issue
CI build passes; new entry appended: '[2026-03-19] Docker CI: multi-arch base image missing arm64 variant → pin --platform linux/amd64'