skills.rest
vh2224
vh2224COMMUNITY
Explore all skills by this author →
GitHub

bad--name

Coordinate multi-LLM engineering workflows with durable markdown state.

1|Updated Jul 13, 2026
Search Tags:#software engineering#code review#autonomous agents#model routing#multi-llm#markdown state
One-click install
npx skills add https://github.com/vh2224/forge-executor --skill bad-name-vh2224
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: bad--name
Source: https://github.com/vh2224/forge-executor/tree/main/packages/pi-coding-agent/test/fixtures/skills/consecutive-hyphens
Command: npx skills add https://github.com/vh2224/forge-executor --skill bad-name-vh2224

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Forge Executor solves the problem of brittle, single-model coding workflows by coordinating planning, execution, review, and escalation across multiple LLM families while keeping all state durable in markdown files.

Core Features & Use Cases

  • Multi-LLM task routing: Assigns planning, execution, review, and advocacy roles to the model best suited for each job.
  • Persistent markdown state: Stores milestones, roadmaps, plans, reviews, knowledge, and ledgers in .gsd/ so work resumes after restarts.
  • Adversarial review workflow: Uses cross-family challenge and defense passes to surface objections before handing unresolved issues back to the user.
  • Use case: A developer launches a milestone, lets the system decompose it into slices, executes the work autonomously, and then resolves review objections with a dedicated fix pass.

Quick Start

Ask Forge Executor to initialize a project milestone, choose suitable models, and carry the work through planning, execution, review, and follow-up in durable markdown state.

Frequently Asked Questions about bad--name

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

FAQPage Schema
How do I run autonomous multi-LLM engineering workflows for software development?▼

You can route software engineering tasks to different models by assigning planning, execution, review, and advocacy roles to specific LLMs, using markdown state to coordinate the autonomous workflow.

What is adversarial cross-family critique in code review?▼

Adversarial cross-family critique is a code review workflow where different LLM families challenge and defend code passes to surface objections before unresolved issues are handed back to the user.

How do I maintain durable state for autonomous coding agents across restarts?▼

You maintain durable state for autonomous coding agents by storing milestones, roadmaps, plans, reviews, and ledgers in markdown files within a `.gsd/` directory, enabling restart-safe task orchestration.

Does model routing improve software engineering workflows compared to single-model approaches?▼

Model routing improves software engineering workflows over single-model approaches by assigning planning, execution, and review to distinct LLMs, preventing brittle workflows and leveraging model strengths for each phase.

Can I use markdown-based project memory for milestone-driven development?▼

Yes, you can use markdown-based project memory for milestone-driven development by storing roadmaps, plans, and knowledge ledgers in markdown, allowing autonomous agents to decompose milestones into slices and execute work durably.

Related Skills

Explore More Software Engineering Skills→
kimasplundkimasplund

qavr-status

Display QAVR memory system status and configuration in a diagnostic report.

Community
Basic
eumemiceumemic

shovel-ready

Drive GitHub shovel-ready labeled issues through TDD to merged pull requests.

Community
Advanced
mkalhitti-cloudmkalhitti-cloud
2

ai-council

Coordinate structured multi-AI debates and synthesize final recommendations.

Community
Advanced
alicicekalicicek
36

tale-mode

Enforces staged planning, adversarial review, and deterministic verification for complex software tasks.

Community
Advanced