Security audit
Security checks for vulnerabilities and agentic risk
This is a disclosed task-orchestration helper, with the main caution that it can continue autopilot work, write project memory, and clean temporary files.
Install this only if you want an agent to manage multi-step OpenClaw subagent workflows with less repeated prompting. Before using it, make sure project memory and task-board updates are acceptable, and preserve any screenshots, logs, or intermediate files you may need before allowing cleanup.
This README is a markdown file, so vague-trigger review applies. The note that the skill is "intentionally generic" provides broad applicability but no concrete invocation phrases, scope boundaries, or exclusion conditions, which can increase the chance of unintended activation in general orchestration contexts.
The skill description is broad enough to auto-activate for many ongoing workflow situations, including multi-round task management and rule updates, without clear exclusions or scope boundaries. In an orchestration skill that can continue dispatching subagents without waiting for user restatement, overbroad triggering increases the chance of unintended autonomous execution, policy drift, or acting on stale context.
The skill explicitly instructs the agent to clean finished sessions and delete temporary artifacts, but it does not require user notice, confirmation, retention rules, or a safety check for whether those artifacts may still be needed. In a continuous autopilot orchestration context, this can lead to silent loss of evidence, logs, screenshots, or intermediate data that may be needed for auditability, rollback, debugging, or user expectations.
No suspicious patterns detected.
