skills.rest
Leonardo-Corte
Leonardo-CorteCOMMUNITY
Explore all skills by this author →
𝕏@Leonardo_go_UpGitHub

maestro-verify

Verify maestro project tasks through pre-ship rituals with PASS/FAIL/HUMAN/BLOCK verdicts.

1|Updated May 26, 2026
Search Tags:#verification#evidence#verdict#cost-budget#architecture-lint#proofmap#AI reviewer
One-click install
npx skills add https://github.com/Leonardo-Corte/local-social-smm-agent --skill maestro-verify-leonardo-corte
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: maestro-verify
Source: https://github.com/Leonardo-Corte/local-social-smm-agent/tree/main/reference/github-agent-orchestration/maestro/skills/bundled/maestro-verify
Command: npx skills add https://github.com/Leonardo-Corte/local-social-smm-agent --skill maestro-verify-leonardo-corte

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

It prevents incomplete or unsafe work from being marked done by enforcing a consistent verification flow that ties planning, evidence, linting, and final verdict semantics together.

Core Features & Use Cases

  • Pre-ship verification ritual: Orchestrates the required sequence of Plan → Implement → Architecture-lint → Verdict request → verdict-based routing (PASS/FAIL/HUMAN/BLOCK).
  • Evidence-backed correctness: Ensures coverage via Plan-Check and ProofMap gates, so acceptance criteria have witness-level evidence before completion.
  • Deterministic safety checks with escalation: Uses architecture-lint for fast structural errors and Trust Verifier for deterministic verdict computation, including cost-budget monitoring and AI Reviewer veto-only policy.

Quick Start

Ask the AI to run the verification loop for a specific task ID by planning and checking the plan file, recording evidence per verification step, and then requesting a final verdict to route the task to ready or back into the fix cycle.

Frequently Asked Questions about maestro-verify

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

FAQPage Schema
How do I enforce pre-ship verification and evidence-backed task completion in my project?▼

Pre-ship verification enforces task completion by applying a canonical ritual that checks plan consistency, architecture-lint findings, proof-map coverage, risk-class policies, and cost-budget monitoring before generating a final verdict.

What is the canonical sequence for generating and checking a task verification verdict?▼

The canonical verification sequence requires planning, implementing, running architecture-lint, recording evidence per acceptance criterion, and requesting a final verdict that routes the task to PASS, FAIL, HUMAN, or BLOCK states.

Do I need a cost-budget monitoring and risk-class policy setup to run architecture-lint and verdict checks?▼

Cost-budget monitoring and risk-class policies are required inputs for the Trust Verifier to compute a deterministic verdict, ensuring unsafe work is blocked and cost constraints are validated before task completion.

How does proof-map coverage validate acceptance criteria before routing a task verdict?▼

Proof-map coverage ensures every acceptance criterion has witness-level evidence recorded during the verification loop, preventing incomplete work from passing the Plan-Check gate and receiving a PASS verdict.

What happens when the AI Reviewer veto policy triggers a BLOCK or FAIL verdict during verification?▼

When the AI Reviewer veto policy triggers a BLOCK or FAIL verdict, the task is routed back into the fix cycle instead of being marked done, requiring another full verification loop iteration.

Can I use this verification protocol for non-trivial tasks that require plan consistency and structural linting?▼

The verification protocol applies to any non-trivial task where plan consistency, architecture-lint findings, and deterministic safety checks with escalation are required before completion.

Related Skills

Explore More Software Engineering Skills→
BermudaLocalsBermudaLocals
1

writing-plans

Generate task-by-task implementation plans with exact file paths and commands.

Community
Advanced
Leonardo-CorteLeonardo-Corte
1

maestro-task

Execute single PR-scoped tasks with claim/block handoffs and verify-driven iteration.

Community
Intermediate
JNZaderJNZader

agent-governance

Define capability models, privilege rings, and anomaly detection for AI agent governance.

Community
Advanced
fagemxfagemx

wusanto-scribe

Standardize acceptance and gate review for AI-generated deliverables.

Community
Advanced