Classify planned actions into ACT, ASK, QUEUE, or ABORT routes.
npx skills add https://github.com/m2ai-portfolio/m2ai-skills-pack --skill judgment-gate
Please help me install this Agent Skill. Skill: judgment-gate Source: https://github.com/m2ai-portfolio/m2ai-skills-pack/tree/main/skills/judgment-gate Command: npx skills add https://github.com/m2ai-portfolio/m2ai-skills-pack --skill judgment-gate
A judgment gate classifies planned actions and routes them to appropriate decision paths to prevent unsafe autonomous execution.
Run judgment-gate on the planned action before proceeding.
High-intent search queries and answers about installing and using this skill.
You can add a safety check by wiring a judgment gate as a PreToolUse hook to classify planned actions and route them to ACT, ASK, QUEUE, or ABORT before execution.
An agent judgment gate evaluates risk-management by applying a four-question rubric covering reversibility, cost boundary, blast radius, and observability to classify and route planned actions.
Yes, a judgment gate can intercept edits and automatically route them to QUEUE if the action violates the cost boundary criteria defined within its four-question safety rubric.
Autonomous execution should route to ABORT instead of ASK when the planned action fails the rubric for reversibility and blast radius, indicating an unsafe and irreversible high-risk operation.
Yes, a PreToolUse hook works with any agent planning tool calls, edits, API requests, or shell commands, applying routing automation to enforce safety before action.
SpillwaveSolutionsConfigure hooks.yaml rules to intercept Claude Code events deterministically.
EntelligentsiaAutomate observe-think-act reasoning loops with termination guardrails.
ricardorocheStandardize Pydantic tool schemas and agent state for multi-agent orchestration.
Kgan01Enforce constitutional safety rules across autonomous agent workflows.
