skills/smithery.ai/aws-coworker-development

aws-coworker-development

Installation
SKILL.md

AWS Coworker Development Guardrails

Scope Definition

This skill applies to ALL files in the aws-coworker-enterprise repository:

Category Examples Gates
CLI Layer (Core) skills/, .claude/agents/, .claude/commands/, config/ Plan approval → Execution authorization → Execute
Server Layer server/ Plan approval → Execution authorization → Execute
Web UI Layer web-ui/ Plan approval → Execution authorization → Execute
Design Docs CLAUDE.md, DESIGN.md, CLAUDE-DEVELOPMENT.md Plan approval → Execution authorization → Execute
User-Facing Docs docs/LESSONS-LEARNED.md, README.md Collaborative editing OK for minor wording; plan + execute gates for structural changes
Archives docs/conversations/ Execution authorization before committing
Tests tests/ Plan approval → Execution authorization → Execute

Cross-layer changes require extra scrutiny. Any change that touches files in more than one layer (CLI + Server, or Server + Web UI) should explicitly justify why both layers need modification and confirm the dependency direction is not violated (Tenet 10).

Plan approval and execution authorization are separate gates. Approving a plan confirms the content is correct. It does NOT authorize the agent to start making changes. See the workflow below.

Installs
1
First Seen
Apr 1, 2026
aws-coworker-development from smithery.ai