Upgrade Openclaw

Upgrade OpenClaw and comprehensively discover new features, config options, hooks, and improvements. Use when: user says "upgrade openclaw", "update openclaw...

This skill appears to do what it says (upgrade and audit OpenClaw) but has two important risks to weigh before installing: 1) It reads live gateway configuration and changelogs and will present every new config key or changelog entry, which can include secrets (API keys, tokens, credentials). 2) It asks you to choose a subagent model and warns 'external providers will receive config data' — if you pick an external model (Anthropic, Deepseek, etc.), your config could be sent off-platform. Recommendations: review SKILL.md fully; if you must run it, pick a local/on-prem model for subagents or ensure no secrets are present in gateway config; back up ~/openclaw and stash important changes manually before running; inspect and remove/rotate any sensitive values in gateway config or limit the tool to dry-run/audit-only modes; confirm you will approve any apply actions (the skill claims it will not apply without explicit approval). Also note minor metadata mismatches (embedded _meta.json version differs from registry version) — not necessarily malicious but worth checking with the skill author if you rely on provenance.

Type: OpenClaw Skill Name: upgrade-openclaw Version: 2.0.0 The skill performs system-level upgrades and configuration audits, which involves reading the full application configuration and potentially sending it to external LLM providers for 'gap analysis'. While the skill is transparent about this behavior in SKILL.md and requires user approval before applying changes, the broad access to system settings and the practice of passing potentially sensitive configuration data to external APIs presents a significant risk of data exposure. Files involved: SKILL.md, settings.json.