pa-ownership

Autonomous task tracking with retry loops and proactive updates. Use when Heleni takes ownership of a task that needs to be tracked, retried on failure, and...

What to check before installing: - Ask the author what the .context file contains and where it comes from. Ensure it does NOT contain secrets you don't want the skill to use or expose. - Confirm how message reactions and notifications are delivered: which messaging connector or credentials will the skill use? Require explicit env vars (messaging API tokens, phone identifiers) rather than hidden .context values. - Review and approve the exact paths it will write to (/opt/ocana/openclaw/workspace/data/pa-tasks.json and memory/whatsapp/*). These files can contain chat history and personally identifiable info; ensure access is limited and backed up if needed. - Approve modification to HEARTBEAT.md (or refuse it). If you do not want the skill to modify global heartbeat behavior, ask the author to provide an optional snippet you can add manually instead of automatic modification. - Test in a sandbox or with a non-production account and monitor what notifications are sent (especially the 'escalate to Netanel' rule). Why I rated it 'suspicious': the skill's functional intent is coherent, but it references and requires runtime data (messaging actions, contact/phone info, .context) without declaring the credentials or explicit permissions. That mismatch — plus its instruction to edit HEARTBEAT.md — increases risk and calls for clarification and least-privilege controls before installation. Additional information that would raise or lower confidence: providing the contents/spec of the .context, declaring required env vars and scopes for messaging, or moving HEARTBEAT integration to an opt-in/manual step would reduce concern; discovery of hidden credential usage or automatic edits to other skills' configs would raise it.

Type: OpenClaw Skill Name: pa-ownership Version: 1.0.0 The pa-ownership skill is a task management system designed for an AI assistant to track, retry, and report on assigned tasks. It uses a local JSON file (/opt/ocana/openclaw/workspace/data/pa-tasks.json) for persistence and includes logic for status updates and heartbeat-driven retries. The code and instructions in SKILL.md are transparent, follow standard OpenClaw patterns for state management, and lack any indicators of malicious intent or unauthorized data access.