Install
openclaw skills install better-every-runBetter Every Run: turn user corrections into preferred future outcomes through a simple /ber-style fix, remember, and report flow.
openclaw skills install better-every-runUse this skill when the user corrects an outcome, states a preference, asks why something keeps failing, or wants an agent to improve future behavior from the current run.
The human path is deliberately small:
/ber fix bad outcome -> desired outcome
/ber remember simple rule
/ber report
The agent runs the bundled local helper in the background and reports the result in chat.
/ber fix agent wrote vague status -> agent gives exact command output and next action
/ber remember use the approved development host for active code work
/ber report
MEMORY.md, AGENTS.md, SOUL.md, or other durable instruction files.For normal chat use, keep the visible flow to one command:
/ber fix bad outcome -> desired outcome
For simple one-line preferences:
/ber remember simple rule
Use /ber report to show what was learned.
Keep helper paths, target-file selection, ledger files, and memory-patch mechanics out of the human-facing answer. The agent may use them internally for deterministic execution, tests, and audits.
Read references/workflow.md for the full workflow.
Read references/report-template.md for expected chat output shape.