skills.rest
tommylower
tommylowerCOMMUNITY
Explore all skills by this author →
Website𝕏@tommylowerGitHub

rams

Route UI review requests to Rams local skill, hosted MCP, or GitHub App surfaces.

2|Updated Apr 15, 2026
Search Tags:#code review#mcp#ui review#design review#github app#rams#data boundary
One-click install
npx skills add https://github.com/tommylower/cortex --skill rams-tommylower
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: rams
Source: https://github.com/tommylower/cortex/tree/main/design/tools/rams
Command: npx skills add https://github.com/tommylower/cortex --skill rams-tommylower

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? It routes requests for the external Rams review service to the correct surface (local checklist, hosted MCP, or GitHub App) while enforcing a data boundary so only approved files leave the machine. ## Core Features & Use Cases - Surface Selection: Choose between the on-device local skill, the metered hosted MCP scored review, and the GitHub App pull-request backstop. - Data Boundary Enforcement: Before any hosted review, state the boundary, list the exact UI files to be sent, and obtain explicit approval while excluding secrets. - Finding Adjudication: Evaluate every Rams finding against the code, rendered interface, project profile, and Studio rules, then fix, decline, or defer it. - Use Case: Before committing a redesigned settings page, ask for a Rams scored review of the changed UI files, adjudicate the findings, and confirm the Cortex-native ship check still passes. ## Quick Start Ask the agent to run a Rams review of the changed UI files before commit and adjudicate each finding against the project rules.

Frequently Asked Questions about rams

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

FAQPage Schema
How do I run a Rams review on my UI files?▼

Choose a surface first: the local skill for a fast checklist pass, the hosted MCP for a scored review via review_files, or the GitHub App for pull-request reviews. For hosted surfaces, list the exact files to send and get explicit approval before running.

What is the difference between Rams local skill, MCP, and GitHub App?▼

The local skill is a stateless checklist that runs in the agent with files staying on-device. The hosted MCP sends only approved files for a scored, metered review. The GitHub App automatically reviews changed UI files on pull requests.

How do I configure the Rams MCP server in Claude Code or Cursor?▼

Create a workspace-scoped API key in Rams settings, then add the Streamable HTTP endpoint https://worker.rams.ai/mcp with an Authorization Bearer header. Never commit the real key; use the client's secret or environment support.

Does Rams send my source files to an external service?▼

The local skill sends nothing. MCP reviews send only files explicitly passed to review_files over TLS, and the GitHub App accesses changed files in enabled repositories. Rams states reviewed source is analyzed in memory and discarded.

When should I use Rams instead of studio-audit or preflight?▼

Use studio-audit as the canonical ship check and preflight for focused static review. Add Rams only as an explicit second opinion after the changed UI is stable and before commit, since Rams findings never override Studio rules.

Related Skills

Explore More Design & Creative Skills→
hackefellerhackefeller
1

jinn-conventions

Document and standardize coding, testing, and release conventions for a project.

Official
Intermediate
learntocloudlearntocloud
738

ship-it

Run pre-commit checks, stage, commit, push, and monitor deployment.

Official
Advanced
involvexinvolvex
2

web-design-guidelines

Review web UI code against Web Interface Guidelines and report violations in file:line format.

Community
Intermediate
try-workstry-works
128

recursive-subagent

Determine subagent delegation availability and enforce handoff contracts during recursive audits.

Community
Advanced