{"skill":{"slug":"clawdgo","displayName":"ClawdGo","summary":"ClawdGo Lobster Cybersecurity Camp. Train one lobster through 3 layers / 12 dimensions with modes W + A-H. Keep onboarding clear, mode boundaries strict, and...","tags":{"agent-security":"1.2.1","chatchess":"1.3.2","cybersecurity":"1.3.2","game":"1.3.2","latest":"1.3.2","phishing":"1.3.2","security-training":"1.2.1","soul-md":"1.2.1"},"stats":{"comments":0,"downloads":319,"installsAllTime":0,"installsCurrent":0,"stars":0,"versions":6},"createdAt":1773481755100,"updatedAt":1777529830518},"latestVersion":{"version":"1.3.2","createdAt":1775535692870,"changelog":"ClawdGo 1.3.2 brings persistent profile/memory architecture, a richer command set, and training flow upgrades:\n\n- Introduced three-layer memory system with persistent profile in `runtime/clawdgo-profile.json`, soul.md anchors, and self-generated scenarios folder.\n- First-run now auto-initializes with 47 training records from seed profile if no profile exists.\n- Expanded trigger words and richer mode/command routing, including new workshop, vaccine, and B-mode explicit/后台/观摩 triggers.\n- Main menu, command card, and mode descriptions updated for new context (\"场景工坊\", \"安全疫苗\", \"H 联网斗虾\" as internal beta).\n- New event sync: session/profile changes now written to JSON on mode/session updates.\n- Removed dependency on legacy soul.md for runtime, aligning with new bootstrapping and profile sync logic.","license":"MIT-0"},"metadata":{"os":null,"systems":null},"owner":{"handle":"nrt2024","userId":"s173z289wxy42yvz6784rejwz183ntn2","displayName":"NRT2024","image":"https://avatars.githubusercontent.com/u/162438792?v=4"},"moderation":{"isSuspicious":true,"isMalwareBlocked":false,"verdict":"suspicious","reasonCodes":["suspicious.exposed_secret_literal"],"summary":"Detected: suspicious.exposed_secret_literal","engineVersion":"v2.4.5","updatedAt":1777529830518}}