skills.rest
bfly123
bfly123COMMUNITY
Explore all skills by this author →
GitHub

node-check

Verify worker node results against acceptance criteria and reject hidden degradation or missing evidence.

3.5k|342|Updated Oct 25, 2025
Search Tags:#verification#acceptance criteria#quality check#evidence review#Multi-Agent#Workflow orchestration#worker node
One-click install
npx skills add https://github.com/bfly123/claude_code_bridge --skill node-check
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: node-check
Source: https://github.com/bfly123/claude_code_bridge/tree/main/docs/plantree/plans/agentic-loop-workflow/drafts/agentroles.ccb_checker/skills/node-check
Command: npx skills add https://github.com/bfly123/claude_code_bridge --skill node-check

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

When a worker agent returns a node result in an agentic workflow, there is no guarantee the work actually met the assigned scope. This Skill provides a structured verification step that catches hidden fallback, silent degradation, scope shrinkage, and missing evidence before results are accepted.

Core Features & Use Cases

  • Structured Verification: Checks a worker result against the task packet reference, assigned node scope, acceptance criteria, and verification expectation.
  • Four-Class Verdict: Returns exactly one of pass, rework_required, blocked, or non_converged, keeping orchestration decisions unambiguous.
  • Evidence-Based Findings: Outputs the check plan, evidence reviewed, findings, and required rework so failures are actionable.
  • Use Case: In a multi-agent CCB workflow, after a worker node claims a task is done, run this check to confirm the deliverable matches the assigned scope and contains real evidence rather than a degraded shortcut.

Quick Start

Verify the worker result for this node against its task packet and acceptance criteria, then return a pass, rework_required, blocked, or non_converged verdict with evidence.

Frequently Asked Questions about node-check

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I verify a worker agent's output in a multi-agent workflow?▼

Run a node check that compares the worker result against the task packet reference, assigned node scope, acceptance criteria, and verification expectation. The check returns one verdict: pass, rework_required, blocked, or non_converged.

What does the node-check skill detect in worker results?▼

It detects hidden fallback, silent degradation, scope shrinkage, and missing evidence in a worker's returned result. Each verdict includes the check plan, evidence reviewed, findings, and any required rework.

What inputs are required to run a node verification check?▼

The check requires five inputs: the worker result, the task packet reference, the assigned node scope, the acceptance criteria, and the verification expectation. All five are needed to produce a reliable verdict.

What is the difference between rework_required and non_converged verdicts?▼

rework_required means the node can be fixed with specific corrective work listed in the findings. non_converged means repeated attempts are not progressing toward the acceptance criteria, signaling the orchestrator should change strategy.

When should node verification run in an agentic loop?▼

Run it immediately after a worker returns a node result and before the orchestrator accepts or aggregates that output. This prevents degraded or incomplete work from propagating into downstream nodes.

Related Skills

Explore More Software Engineering Skills→
agentscope-aiagentscope-ai
5.5k

teamharness-team-coordination

Plans task boundaries, ownership, and DAG or Loop execution modes for multi-agent team coordination.

Official
Intermediate
breenglesbreengles
4

todo-implement

Orchestrate TODO item implementation with implementer and reviewer subagents.

Community
Advanced
rbtkhnrbtkhn

skill-jiang

Predict future lecture topics from historical prefixes with scored reveal cycles.

Community
Advanced
InteligentsensingsolutionsInteligentsensingsolutions
1

e2e-testing

Validate end-to-end user workflows through the UI with Playwright.

Official
Intermediate