跳至正文
OpenAgentSkill
浏览已开放 · 目录记录日期 2026-11-21 · 自动更新和账户服务暂时关闭。
法律与合规

rootnode-critic-gate

Independent re-derivation gate for proposed changes during autonomous execution. Evaluates a change (code diff, config edit, engine evolution, schema change) against the work's authority matrix and a 4-check protocol: invariant compliance, scope authorization, detection narrowness, and regression risk. Returns structured JSON with pass/fail per check, an overall verdict (APPROVE / REQUEST_CHANGES / REJECT), and blockers. Profile-driven thresholds (strict for unattended runs, lenient for desk supervision). Use when an autonomous agent proposes an engine change, when reviewing a Claude Code-authored modification before merge, or when an unattended profile requires independent re-derivation. Trigger on: "review/approve this proposed change," "critic-gate this," "is this change safe," "is this safe to merge," "should I let this land," "run the critic." Do NOT use for handoff-readiness checks (use rootnode-handoff-trigger-check), general code review, or design-time correctness. Authority ma

rootnode · 已有静态检查记录

安装与使用

运行命令前先阅读来源与依赖说明。复制命令不会自动安装或执行。

npx skills add drayline/rootnode-skills --skill rootnode-critic-gate

技能说明


name: rootnode-critic-gate description: >- Independent re-derivation gate for proposed changes during autonomous execution. Evaluates a change (code diff, config edit, engine evolution, schema change) against the work's authority matrix and a 4-check protocol: invariant compliance, scope authorization, detection narrowness, and regression risk. Returns structured JSON with pass/fail per check, an overall verdict (APPROVE / REQUEST_CHANGES / REJECT), and blockers. Profile-driven thresholds (strict for unattended runs, lenient for desk supervision). Use when an autonomous agent proposes an engine change, when reviewing a Claude Code-authored modification before merge, or when an unattended profile requires independent re-derivation. Trigger on: "review/approve this proposed change," "critic-gate this," "is this change safe," "is this safe to merge," "should I let this land," "run the critic." Do NOT use for handoff-readiness checks (use rootnode-handoff-trigger-check), general code review, or design-time correctness. Authority matrix required. license: Apache-2.0 metadata: author: rootnode version: "4.0.0" original-source: "root.node design 2026 — distilled from production engine evolution patterns" companion-files: "schema/profile.schema.json, profiles/lenient.json, profiles/strict.json, profiles/balanced.json, references/checks-detailed.md, references/severity-coverage.md, references/troubleshooting.md" changelog: "1.0.2 (2026-05-05): Activation precision tuning. Trigger list rebalanced — 'review this proposed change' + 'approve this diff' consolidated to 'review/approve this proposed change' (saves char budget); 'should this land' replaced with 'should I let this land' (1st-person symptom phrasing); 'is this safe to merge' added as new symptom trigger (merge framing covers user-vocabulary surface that change/diff framing misses). Description grew 996 → 1014 chars (10-char headroom remaining). No methodology, schema, or workflow changes; behavior identical to 1.0.1 once activated. 1.0.1 (2026-05-01): Structural patch to align with rootnode authoring convention. Added references/ folder with three on-demand-loaded files (checks-detailed, severity-coverage, troubleshooting). SKILL.md body slimmed via content extraction; behavior identical to 1.0. Deployment target: Claude Code (CC) side of CP/CC split." discipline_post: phase-30

Critic Gate

Calibration: Tier 2 (Sonnet-graceful) - designed and tested against Opus 5 + Sonnet 5 dual-primary. Runs correctly on Sonnet 4.6 (legacy-graceful) with slightly less depth on multi-dimensional analysis; on Haiku 4.5 with extended thinking; fallback-graceful on Opus 4.8. Effort default is high on Opus 5 and Sonnet 5; step down to medium for cost-sensitive runs where evals show quality holds. See repository README for model compatibility.

Independent re-derivation of proposed changes during autonomous execution. Returns a structured verdict before any change lands. The load-bearing safety mechanism for autonomous engine evolution — without it, multi-cycle drift is guaranteed.

The premise: the agent that proposes a change cannot be the agent that approves it. When Claude Code is iterating on an engine, every proposed change has an author who has reasoned themselves into the change's necessity. That author's judgment is exactly what the change is grounded in — making them an unreliable evaluator of the change's safety. The critic gate runs on a fresh evaluator with no commitment to the change, applying narrow, named checks against the work's invariants.

This Skill is the rootnode runtime layer's analog to a "Critic Agent" role — an independent re-derivation pass over proposed changes. Generalized for portability: the authority matrix is supplied as input, not baked in.

Important

The critic does not re-derive the diagnosis. It does not verify that the bug exists or that the proposed fix would solve it. Those are the change author's responsibility. The critic checks one thing: would applying this change violate any invariant, exceed any authorization, or introduce regression risk that wasn't accepted in the work's design? The diagnosis can be wrong; that's a verification problem, not a critic problem.

Authority matrix is mandatory input. The critic cannot evaluate compliance against rules it doesn't have. If no authority matrix is provided in work_context, the gate halts with evidence_not_provided rather than fabricating rules. This is the most common failure mode to design against.

Structured JSON output is the deliverable. Same shape as rootnode-handoff-trigger-check (if available; the shape is documented inline below regardless). An orchestrator should be able to parse the verdict and act on it without an LLM round-trip.

Three verdicts, not two. Unlike handoff-trigger-check (PASS/FAIL), critic gate has three: APPROVE (proceed), REQUEST_CHANGES (specific issues to address; change author re-proposes), REJECT (proposed change violates non-negotiable constraints; do not re-propose without scope-authorization escalation). The middle verdict matters — most real critic findings are fixable, not fatal.

Profile thresholds calibrate strictness. Same profile pattern as handoff-trigger-check. Sleeping profiles run all checks at strictest thresholds; desk profiles allow REQUEST_CHANGES on minor findings to be auto-approved if the change author resolves them in the same turn.

Severity coverage requirement. Every severity (info, minor, major, blocker) must be routed by the profile to exactly one of auto_approve_on, request_changes_threshold, or reject_on. Unrouted severities create ambiguous runtime behavior. The Skill validates coverage on profile load and rejects misconfigured profiles before any change is evaluated. See references/severity-coverage.md for the full routing rules and example profile patterns.


The 4 Checks — Overview

阅读完整来源 ↗