Proactive Agent 3.1.0

This skill is largely what it claims to be (a set of policies and local files to make an agent proactive), but there are a few red flags to consider before installing or enabling it on a live agent: - Source provenance: The package has no homepage and the source is 'unknown' — prefer skills from known authors or inspect everything before trusting. - Conflicting guidance: Some files say 'Don't ask permission. Just do it.' while others require explicit human approval for external actions. Decide which behaviour you want and reconcile the files before enabling autonomous operation. - Test offline/sandbox: Run the included scripts and a copy of the SKILL files in an isolated workspace (no network access, no real credentials) to see what the agent would do. - Check .credentials and .gitignore: Verify no real secrets are present in workspace files, ensure .credentials is gitignored and permissions are correct (the audit script checks this). - Be careful with network and agent-to-agent features: The documentation suggests optional connections to WhatsApp/Telegram and spawning agents; do not provide tokens or enable network connectivity until you review/limit those behaviors. - Review and harden heartbeats/autonomy: If you allow autonomous invocation, restrict or gate actions that send data externally, publish, or delete things unless you explicitly approve them. If you want, I can: - Highlight the exact lines that encourage acting without permission or connecting externally, - Produce a minimal checklist of edits to make the skill safer (e.g., remove 'Don't ask permission' and add explicit gating), or - Run a focused audit of the script contents and all files and summarize actionable changes.