build-loop

Apply planned repository changes by reading BUILD context files and logging each attempt to build-log.md.

1|Updated Mar 16, 2026
One-click install
npx skills add https://github.com/jackson2787/torque-coding --skill build-loop-jackson2787
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: build-loop
Source: https://github.com/jackson2787/torque-coding/tree/main/skills/state-machine/build-loop
Command: npx skills add https://github.com/jackson2787/torque-coding --skill build-loop-jackson2787

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Builds are automated by applying a planned set of changes to a repository without requiring manual codebase exploration, reducing drift and human error.

Core Features & Use Cases

  • Zero-exploration execution: BUILD applies plan_context-defined changes strictly per the plan, logging each attempt.
  • Doctrinal compliance: reads constitution.md, operational-context.md, plan.md, plan_context.md, and rules/execution-discipline.md to guide execution and escalation when gaps are detected.
  • Escalation and governance: supports escalation to PLAN-CONTEXTUALIZE or escalate upon hard constraints or missing approval.
  • Build logging and governance: maintains a build-log.md with per-attempt details and outcomes; enforces hard and soft caps.

Quick Start

Run the build-loop using the current plan and plan_context to execute changes and log the attempt.

Frequently Asked Questions about build-loop

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

FAQPage Schema
How do I automate applying a planned set of changes to a repository without exploring the codebase?

To automate applying planned changes without exploring the codebase, use a deterministic plan execution tool that reads plan.md, plan_context.md, and rules/execution-discipline.md to apply changes strictly per the plan. This enforces a zero-exploration happy path and outputs git-diff style changes.

What is deterministic plan execution for code changes?

Deterministic plan execution for code changes is an automated process that reads an approved plan and its context to apply repository changes without exploration. It logs each attempt to build-log.md, enforces per-attempt hard caps, and escalates when the plan pack is incomplete.

How do I enforce governance and hard caps during automated code change execution?

To enforce governance and hard caps during automated code change execution, the execution process reads constitution.md and operational-context.md to guide actions. It maintains a build-log.md with per-attempt details and escalates to PLAN-CONTEXTUALIZE upon hitting hard constraints or missing approval.

Do I need an approved plan and context files to start automated build execution?

Yes, you need an approved plan to start automated build execution. The process requires plan.md, plan_context.md, constitution.md, operational-context.md, and rules/execution-discipline.md to guide execution, and it will escalate if the pack is incomplete or missing required approval.

When should I escalate instead of continuing automated plan execution?

You should escalate instead of continuing automated plan execution when the plan pack is incomplete, when hard constraints are hit, or when approval is missing. The execution process escalates to PLAN-CONTEXTUALIZE upon detecting these gaps to ensure doctrinal compliance.

What format do automated build execution tools output for repository changes?

Automated build execution tools output repository changes in a git-diff style format as the result. This provides a clear, standard representation of the applied changes without requiring manual codebase exploration during the build process.