菜单
Use when the user asks to install or configure ai-sec for coding agents (for example: Claude hooks, Codex wrapper, agent-first setup, local integration install). This skill installs and verifies ai-sec integrations for Claude Code and Codex CLI.
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 | ai-sec-bootstrap |
| description | Use when the user asks to install or configure ai-sec for coding agents (for example: Claude hooks, Codex wrapper, agent-first setup, local integration install). This skill installs and verifies ai-sec integrations for Claude Code and Codex CLI. |
Use this skill to install ai-sec integrations for agent CLIs with minimal manual steps.
Run when the user asks to set up ai-sec in Claude Code or Codex CLI.
scripts/install_integrations.sh.~/.claude/hooks/claude-ai-sec-hook.py~/.claude/settings.json contains UserPromptSubmit and PreToolUse hook entries.~/.local/bin/codex-ai-secInstall both:
./scripts/install_integrations.sh
Install only Claude:
./scripts/install_integrations.sh --claude-only
Install only Codex wrapper:
./scripts/install_integrations.sh --codex-only
AI_SEC_GATEWAY_URLAI_SEC_BEARER_TOKEN (or SERVICE_API_TOKEN)Optional:
AI_SEC_FAIL_CLOSED=1 (recommended for strict environments)Use when the user wants to integrate ai-sec into an OpenClaw bot (for example: "add ai-sec to openclaw", "guard openclaw tools", "openclaw prompt injection defense", "scaffold openclaw security middleware"). This skill scaffolds OpenClaw guard code using @codegrammer/ai-sec-openclaw-adapter with autonomous mode by default and optional human approval mode.
Use when a user asks to run prompts or tool actions through ai-sec before execution (for example: "guard this prompt", "gate tool calls", "run safely", "agent-first security", "preflight this task"). This skill performs ai-sec preflight gating and enforces allow/review/block exit codes.
Use when the user asks to operate, validate, or monitor ai-sec (for example: check gateway health, run red-team suite, inspect security events, or validate policy behavior). This skill runs deterministic ai-sec operations via helper scripts.