skills.rest
Maxcogar
MaxcogarCOMMUNITY
Explore all skills by this author →
GitHub

project-contractor

Survey codebases with parallel subagents to generate SOURCE-OF-TRUTH.md.

Updated Jan 5, 2026
Search Tags:#architecture#codebase#verification#process#survey#guardrails#source-of-truth
One-click install
npx skills add https://github.com/Maxcogar/agent-armory --skill project-contractor
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: project-contractor
Source: https://github.com/Maxcogar/agent-armory/tree/main/skills/project-contractor
Command: npx skills add https://github.com/Maxcogar/agent-armory --skill project-contractor

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

Claude often proceeds without full project understanding, risking incorrect changes. This skill ensures a comprehensive survey that builds a canonical SOURCE-OF-TRUTH.md before any work.

Core Features & Use Cases

  • Parallel subagent surveys to cover architecture, connections, operations, and state
  • Generates SOURCE-OF-TRUTH.md as the canonical reference for the project
  • Prevents assumptions by requiring explicit understanding before acting

Quick Start

Run a full Phase 1 survey to generate SOURCE-OF-TRUTH.md and validate with the user before proceeding.

Frequently Asked Questions about project-contractor

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

FAQPage Schema
How do I survey an unfamiliar codebase before making changes?▼

A codebase survey builds a single source of truth by using parallel subagents to inspect architecture, connections, operations, and state. It outputs SOURCE-OF-TRUTH.md with canonical commands to prevent incorrect changes.

Why does Claude make incorrect changes to my project architecture?▼

Incorrect changes happen when acting without full project understanding. A comprehensive codebase survey generates SOURCE-OF-TRUTH.md, requiring explicit verification of architecture and state before proceeding with modifications.

What is the best way to create a source of truth for an existing codebase?▼

The best way is running a Phase 1 survey that coordinates parallel subagents across architecture, connections, operations, and state. This generates SOURCE-OF-TRUTH.md with canonical commands and guardrails as a canonical reference.

Can I use this codebase survey process when re-engaging after context loss?▼

Yes, this codebase survey process is explicitly applicable when re-engaging after context loss or starting new projects. It rebuilds project understanding by generating a fresh SOURCE-OF-TRUTH.md file validated with the user.

How do I generate guardrails and canonical commands for my source code?▼

You generate guardrails and canonical commands by running a full Phase 1 survey of the codebase. The parallel subagents analyze architecture and state to output these guardrails within the SOURCE-OF-TRUTH.md file.

Related Skills

Explore More Software Engineering Skills→
OzeroHAXOzeroHAX
32

docs-project-brief

Define project context, problem statement, and MVP boundaries.

Community
Basic
AztecProtocolAztecProtocol
449

updating-changelog

Generate changelog drafts by analyzing git diffs between next and HEAD.

Official
Intermediate
Marshall-HallenbeckMarshall-Hallenbeck
5

fix-tests

Diagnose and repair failing test suites across multiple language ecosystems.

Community
Advanced
tyronerosstyroneross
1

build-loop:architecture-scan

Scan project architecture components and connections for health issues.

Community
Advanced