Peer: Claude Skill Makes AI Answer Like a Colleague

A new open-source skill called Peer transforms Claude's responses to sound like messages from a human colleague rather than a formal AI assistant. Developer Andy Kear released the technique on GitHub today, with early testing showing users could not reliably distinguish Peer-enhanced responses from actual coworker communications. The skill addresses a common complaint about AI assistants: they sound robotic, overly polite, and distinctly non-human.
How Peer Reshapes Claude Responses
Peer works through carefully engineered system prompts that modify Claude's default communication patterns. Instead of producing formal, structured answers with clear topic sentences and corporate politeness, the skill guides Claude to adopt the casual shorthand, incomplete thoughts, and natural flow typical of workplace instant messaging. The result is responses that feel conversational and peer-to-peer rather than assistant-to-user.
The technique does not require custom model training or API modifications. Users can implement Peer by incorporating specific prompt instructions into their Claude interactions. This accessibility makes it immediately usable for anyone with Claude access, from individual developers to teams seeking more natural AI collaboration.
Blind Testing Results
Kear subjected Peer to blind testing where participants reviewed message threads without knowing which responses came from Claude with Peer applied versus actual human colleagues. According to the GitHub documentation, testers struggled to identify the AI-generated content consistently. This validation suggests Peer successfully bridges the uncanny valley that often separates AI communication from human conversation.
The testing methodology focused on realistic workplace scenarios: code review comments, project update messages, technical troubleshooting discussions, and brainstorming threads. In each context, Peer-enhanced responses maintained technical accuracy while adopting the informal markers of human workplace communication.
Release Date and Availability
Officially released on November 17, 2026, Peer is available as an open-source project on GitHub. The repository includes implementation instructions, example prompts, and documentation explaining the linguistic principles behind the conversational transformation. Developers can fork the project, adapt the prompts to their specific use cases, or integrate the technique into existing Claude workflows.
The Hacker News community responded with interest, noting the skill's potential for reducing AI detection in collaborative environments and improving user comfort during extended AI interactions. Some users highlighted applications in customer support, technical documentation, and internal team communication tools.
Technical Implementation
Peer operates through prompt engineering rather than model fine-tuning. The core technique involves:
- Instructing Claude to avoid formal greeting patterns and sign-offs
- Encouraging sentence fragments, casual language, and natural speech rhythms
- Removing excessive politeness markers like "I'd be happy to" or "Please let me know"
- Mimicking the cognitive shortcuts humans use in rapid workplace communication
This approach maintains Claude's reasoning capabilities and factual accuracy while stripping away the linguistic markers that signal artificial intelligence. The skill does not compromise response quality or introduce errors; it simply repackages accurate information in human-sounding delivery.
What This Means
Peer represents a shift in how developers think about AI integration into collaborative workflows. Rather than treating AI as a distinct entity with its own communication style, the skill enables seamless blending into existing team dynamics. For organizations deploying Claude in customer-facing or internal communication roles, Peer offers a path to more natural interactions without custom model development. The open-source release ensures rapid experimentation and adaptation across diverse use cases, potentially influencing how future AI assistants approach conversational design by default.
on Emergent today





