Security audit
Security checks for vulnerabilities and agentic risk
This is a markdown-only workflow for organizing agent setup hardening and task handoffs, with no bundled code, network behavior, or hidden execution.
Install this only if you want an agent workflow that may review and adjust workspace instruction or memory files. Check proposed memory or agent-guidance changes before accepting them, especially in shared or production repositories.
The phrase 'Use this skill as the top-level entry point when you need a reliable, safety-first agent operating system' is an open-ended activation description without concrete scope boundaries or exclusion conditions. This could cause the skill to be invoked for a wide range of ordinary agent tasks rather than a narrowly defined scenario.
Conditions such as 'Existing workspace feels inconsistent/noisy' and 'Complex multi-step tasks with quality drift' rely on subjective judgment and do not clearly distinguish activation from non-activation cases. The absence of explicit thresholds or examples increases the chance of unintended or inconsistent invocation.
No suspicious patterns detected.
