What is Superego for Claude Code?

It is a skill that adds a 'thinking' layer to Claude, allowing it to evaluate its own plans and your code for strategic alignment, focus, and potential risks.

Does it support non-coding tasks?

Yes, Superego includes specialized prompts for technical writing and learning to ensure quality and coach-like feedback across different domains.

What happens if Superego finds a concern?

If the evaluation returns 'has_concerns: true', Claude is instructed to stop, show you the feedback verbatim, and wait for your approval before proceeding.

When should I use the $superego command?

You should invoke it at critical decision points, such as before committing to a complex architectural change, when choosing between alternatives, or before finalizing a non-trivial feature.

How do I install Superego?

Simply run '$superego init' within your project. This will install the necessary binaries, set up the configuration directory, and update your AGENTS.md file with guidance.

Superego Metacognitive Oversight

Name: Superego Metacognitive Oversight
Author: open-horizon-labs

byopen-horizon-labs

•

Productivity & Workflow

Provides metacognitive oversight to evaluate plan alignment, focus, and proportionality during development tasks.

About

Superego is a specialized Claude Code skill designed to act as a metacognitive advisor, helping developers maintain strategic focus and avoid common pitfalls. By invoking a 'stop and think' mechanism, it evaluates the current conversation context for alignment with project goals, proportionality of effort, and potential risks before non-trivial implementations. It features context-aware modes for software development, technical writing, and learning, offering on-demand reviews and automated guardrails at critical decision points to ensure high-quality, intentional output.

Key Features

Customizable prompt profiles and local configuration
Context-specific prompts for coding, writing, and learning
Automated 'has_concerns' feedback loop to prevent errors
Metacognitive evaluation to catch strategic mistakes
On-demand review of staged changes and pull requests
11 GitHub stars

Use Cases

Reviewing git diffs or pull requests for architectural alignment
Evaluating a complex implementation plan before starting work
Maintaining focus and proportionality in long-running development tasks

About

Key Features

Customizable prompt profiles and local configuration
Context-specific prompts for coding, writing, and learning
Automated 'has_concerns' feedback loop to prevent errors
Metacognitive evaluation to catch strategic mistakes
On-demand review of staged changes and pull requests
11 GitHub stars

Use Cases

Reviewing git diffs or pull requests for architectural alignment
Evaluating a complex implementation plan before starting work
Maintaining focus and proportionality in long-running development tasks