skills/architect-implementation-documentation/assets/coding-agent-first-read-template.md

# Coding-Agent First-Read

## Control and authority

Project / suite ID:  
First-read version/status:  
Implementation scope:  
Current phase:  
Current gate:  
Authorized by / evidence:  

## Project expert matrix

| Phase/task | Active skill | Lead | Support | Independent review | Decision classes |
|---|---|---|---|---|---|

## Current task execution contract

Task ID:  
Objective:  
Fixed / targets / bounds:  
Class-2 discretion:  
Class-3 discretion:  
Representation strategy:  
Local checkpoint:  
Pass condition:  
Rollback:  

## Authoritative source manifest and usage

| Read order | Artifact + version | Status | Usage class / permitted sections | Purpose |
|---|---|---|---|---|

## Precedence and conflict handling

## Permitted changes

## Protected and prohibited changes

## Open decisions and blocked areas

## Last approved checkpoint and perceptual status

## Logged Class-2 decisions

## Implementation sequence

## Verification commands and evidence

## Acceptance criteria and definition of done

## Stop and escalation conditions

Do not make an unauthorized Class-1 decision or resolve a controlled-source conflict. Exercise Class-2 judgment only inside the named role, target, bounds, evidence, checkpoint, and review contract. Stop the affected area and report the exact source gap.