Use when you need to screen inbound untrusted text before it reaches a primary model — detecting prompt injection by its behavioral effect on a sacrificial canary model, not just pattern matching — and want a block/flag/pass routing decision plus explicit behavioral-coverage state. Inbound preflight sensor, not a guarantee.
安装
npx skills add hermes-labs-ai/little-canary --skill little-canarynpx skills add hermes-labs-ai/little-canary --skill little-canary -a claude-code -gnpx skills add hermes-labs-ai/little-canary --skill little-canary -a codex -ggemini skills install https://github.com/hermes-labs-ai/little-canary.git --path .agents/skills/little-canarygh skill install hermes-labs-ai/little-canary little-canarygit clone --depth 1 https://github.com/hermes-labs-ai/little-canary /tmp/hermes-labs-ai__little-canary
mkdir -p ~/.claude/skills ~/.agents/skills
cp -R /tmp/hermes-labs-ai__little-canary/.agents/skills/little-canary ~/.claude/skills/little-canary # Claude Code
cp -R /tmp/hermes-labs-ai__little-canary/.agents/skills/little-canary ~/.agents/skills/little-canary # Codex, Cursor, Gemini CLI, Copilot, OpenCode