菜单
Use this for securing AI-agent workspaces themselves, including instruction files, MCP setups, memory, external content, prompt-injection surfaces, and least-privilege configuration.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
| name | bosskuai-agent-security-hardening |
| description | Use this for securing AI-agent workspaces themselves, including instruction files, MCP setups, memory, external content, prompt-injection surfaces, and least-privilege configuration. |
Use this for securing AI-agent workspaces themselves, including instruction files, MCP setups, memory, external content, prompt-injection surfaces, and least-privilege configuration.
Open ../../references/playbooks/bosskuai-agent-security-hardening-playbook.md only when the task needs detailed framework choices, longer checklists, examples, or implementation depth.
Before finalizing, check:
Use this for Hyperframes video composition work, including HTML-based scene authoring, preview/render workflows, CLI commands, agent-friendly composition updates, and GSAP-assisted animation inside Hyperframes projects.
Use when a Codex workspace needs BosskuAI routing, repo-local memory, skill selection, or quality gates without loading the full toolkit.
Use this for cloud/server/API/model cost control, token budgets, queue sizing, storage costs, vendor spend, and unit economics.
Use this for support SOPs, onboarding, ticket triage, customer success playbooks, retention, churn reduction, and feedback loops.
Use this for agent eval design, routing tests, retrieval tests, LLM quality cases, regression harnesses, scorecards, and continuous agent improvement.
Use this for Malaysia PDPA-aware privacy review, data minimization, consent, retention, user rights, vendor processors, and privacy-safe SaaS operations.
