gsd-plan-review-convergence

Iteratively regenerate plans until external reviewers report no HIGH concerns.

Updated May 21, 2026
One-click install
npx skills add https://github.com/jedmamosto/m-and-ms --skill gsd-plan-review-convergence
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: gsd-plan-review-convergence
Source: https://github.com/jedmamosto/m-and-ms/tree/main/.agents/skills/gsd-plan-review-convergence
Command: npx skills add https://github.com/jedmamosto/m-and-ms --skill gsd-plan-review-convergence

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill removes the uncertainty of multi-AI planning by repeatedly reviewing generated plans for HIGH concerns and re-planning until the plan converges on an acceptable risk level.

Core Features & Use Cases

  • Cross-AI review gate: Runs external AI reviewers (e.g., Codex, Gemini, Claude, OpenCode, and optional local servers) against plans produced by the existing planner flow.
  • Iterative replan until safe: Detects remaining HIGH concerns and triggers a replan using review feedback until none remain or a maximum cycle limit is hit.
  • Workflow orchestration layer: Operates as an outer revision controller around existing Skills, delegating execution to the planner and reviewer Skills while handling loop control, gating, and escalation.

Quick Start

Run gsd-plan-review-convergence with a phase number and enable the plan_review_convergence workflow gate first so the revision loop can execute.

Frequently Asked Questions about gsd-plan-review-convergence

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

FAQPage Schema
How do I automate AI plan review to fix high-risk concerns?

To automate AI plan review for high-risk concerns, you need a gated revision workflow that repeatedly detects HIGH issues in generated plans and triggers replanning using review feedback until none remain or a maximum cycle limit is reached.

What is cross-AI workflow orchestration for plan validation?

Cross-AI workflow orchestration for plan validation operates as an outer revision controller that delegates execution to existing planner and reviewer Skills in isolated Agents, handling loop control, risk gating, and escalation across multiple models.

Can I use external CLI reviewers like Codex and Gemini to gate generated plans?

Yes, external CLI reviewers like Codex, Gemini, Claude, and OpenCode can gate generated plans by running against plans produced by the existing planner flow to identify remaining HIGH concerns before approval.

How do I set up a revision loop for multi-model plan convergence?

Setting up a revision loop for multi-model plan convergence requires enabling the plan_review_convergence workflow gate first, then running the process with a phase number so the gated revision loop can execute.

What happens when iterative replanning cannot resolve high concerns?

When iterative replanning cannot resolve HIGH concerns, the gated revision workflow stops after hitting a maximum cycle limit, ensuring the convergence process terminates gracefully instead of looping indefinitely on unresolvable risk issues.