7cubit
UserMulti-model defensive security review skill for Claude Code — fan a target across an ensemble of frontier models and reconcile their findings. Blue-team only; bring your own keys.
Categories
Indexed Skills (2)
cubersecurity
Run a defensive security review by fanning a target out across a multi-model ensemble (organizer + worker/decision tier) and consolidating the findings. Use this whenever the user wants a security audit, vulnerability review, secure code review, PR/diff security pass, infrastructure or config hardening review, dependency / supply-chain audit, threat model, or incident triage — on code, infrastructure, or systems they own or are explicitly authorized to test. Trigger it even when the user just says things like "is this safe", "review this for vulns", "harden this", "what could go wrong with this design", or "check these deps" without using the word "security". Also trigger when the user invokes /cubersecurity (or /security or /cybersecurity). Do NOT use this to produce weaponized exploit code, malware, or to attack systems the user does not control �� it identifies, assesses, and remediates, it does not weaponize.
cybersecurity
Run a defensive security review by fanning a target out across a multi-model ensemble (organizer + worker/decision tier) and consolidating the findings. Use this whenever the user wants a security audit, vulnerability review, secure code review, PR/diff security pass, infrastructure or config hardening review, dependency / supply-chain audit, threat model, or incident triage — on code, infrastructure, or systems they own or are explicitly authorized to test. Trigger it even when the user just says things like "is this safe", "review this for vulns", "harden this", "what could go wrong with this design", or "check these deps" without using the word "security". Also trigger when the user invokes /cybersecurity (or /security). Do NOT use this to produce weaponized exploit code, malware, or to attack systems the user does not control — it identifies, assesses, and remediates, it does not weaponize.
Bio shown is the top-scored skill's repo description as a fallback — real GitHub bios land in a future update.