BotLearn HealthCheck

This skill is an instruction-only health inspector that will autonomously run local scripts and read many files under your OpenClaw installation (logs, cron, identity, models, openclaw.json, workspace files). That behavior is consistent with a deep health check, but (1) the registry metadata does not fully declare those config paths, and (2) the skill explicitly tells the agent to act without asking you. Before installing or enabling autonomous use: - Verify the skill's source/owner and prefer installing only from a trusted registry or repository. The published metadata shows an unknown source. - Inspect the referenced docs and scripts (setup.md, data_collect.md, check_*.md, openclaw_knowledge.md) to confirm exactly what will be executed and what files will be read. Because this is instruction-only, those documents are the only place containing operational details. - If you have secrets or production systems under $OPENCLAW_HOME, run this skill first in a sandbox/non-production copy of OPENCLAW_HOME or set OPENCLAW_HOME to a safe test folder. - Consider disabling autonomous invocation (or require explicit user confirmation) so the agent cannot run the full collection without your approval. - Monitor outputs for accidental secret exposure (credential strings, tokens, private keys) and rotate any credentials if sensitive data is exposed during a test run. Given the metadata discrepancies and breadth of file access, proceed only after you review the detailed collection/check files or test in an isolated environment.