brutal-critic

Use this agent to get an uncompromising quality evaluation of code, architecture, or technical solutions. It researches best practices online, evaluates against an 8/10 minimum standard, and delivers a structured PASS/FAIL verdict with cited sources. <example> Context: The user wants a code review with teeth. user: "Review the authentication module in src/auth/" assistant: "I'll use the brutal-critic-code agent to evaluate your authentication module against industry best practices." <commentary> The user wants quality evaluation of a specific code module — direct trigger for the brutal-critic-code agent. </commentary> </example> <example> Context: The user just finished implementing a feature and wants honest feedback. user: "Be brutally honest — is this API design any good?" assistant: "I'll launch the brutal-critic-code agent to evaluate your API design against established patterns and standards." <commentary> Explicit request for harsh evaluation of code — exactly what the brutal-critic-code agent is built for. </commentary> </example> <example> Context: The user is choosing between architectural approaches. user: "Evaluate whether our current error handling strategy is production-ready." assistant: "I'll use the brutal-critic-code agent to assess your error handling against production-grade standards with cited best practices." <commentary> Production-readiness assessment of code requires rigorous evaluation — use the brutal-critic-code agent. </commentary> </example> <example> Context: The user wants to validate code before a release. user: "Run a quality gate check on the payment processing service before we ship." assistant: "I'll deploy the brutal-critic-code agent to run a full quality evaluation on the payment service — nothing ships below 8/10." <commentary> Pre-release quality gate on code — the brutal-critic-code agent enforces the 8/10 minimum standard. </commentary> </example> <example> Context: Another agent produced code that needs validation. assistant: "I've implemented the new caching layer." user: "Now tear it apart — what's wrong with it?" assistant: "I'll use the brutal-critic-code agent to find every weakness in the caching implementation." <commentary> The user explicitly wants destructive criticism of recently written code — perfect use case for the brutal-critic-code agent. </commentary> </example>

Use this agent to get an uncompromising quality evaluation of any non-code text — technical documentation, articles, emails, messages, ideas, or arguments. It auto-detects the text type, evaluates against an 8/10 minimum standard, delivers a structured PASS/FAIL verdict, and provides an improved rewrite of weak sections. <example> Context: The user wrote an architecture document and wants honest feedback. user: "Review this ADR I wrote for the migration to microservices." assistant: "I'll use the brutal-critic-text agent to evaluate your architecture decision record against technical writing best practices." <commentary> Technical documentation evaluation — the text agent auto-detects this as technical writing and loads the appropriate framework. </commentary> </example> <example> Context: The user drafted an email and wants it checked. user: "Is this email to the client professional enough?" assistant: "I'll launch the brutal-critic-text agent to evaluate your email for clarity, tone, and professionalism." <commentary> Communication evaluation — the text agent detects this as a professional message and evaluates accordingly. </commentary> </example> <example> Context: The user wrote a blog post or article. user: "Tear apart this article I wrote about distributed systems." assistant: "I'll use the brutal-critic-text agent to evaluate your article for structure, clarity, argumentation, and engagement." <commentary> General writing evaluation — explicit request for harsh feedback on written content. </commentary> </example> <example> Context: The user has an idea and wants it stress-tested. user: "I'm thinking we should switch from REST to GraphQL for all our APIs. Is this reasoning solid?" assistant: "I'll deploy the brutal-critic-text agent to evaluate the logical coherence, evidence, and feasibility of your argument." <commentary> Thought/idea evaluation — the user wants their reasoning critically examined, not code reviewed. </commentary> </example> <example> Context: The user wants to check if their text is grammatically correct. user: "Check if this text reads well in English." assistant: "I'll use the brutal-critic-text agent to evaluate your text for grammar, clarity, and readability." <commentary> General English quality check — the text agent evaluates language mechanics and style. </commentary> </example> <example> Context: The user created architecture diagrams and wants feedback. user: "Evaluate these architecture diagrams for the new payment system." assistant: "I'll use the brutal-critic-text agent to evaluate your architecture diagrams for accuracy, completeness, and clarity." <commentary> Architecture diagram evaluation — diagrams are documentation, not code. The text agent handles this as technical writing. </commentary> </example>

Structured code evaluation framework used by the brutal-critic-code agent. Defines the 1-10 rating scale with 8/10 minimum pass threshold, 9 evaluation categories (correctness, error handling, architecture, readability, security, performance, testing, linting & formatting, API design) with pass/fail criteria, the report template, and scoring rules including weakest-link overall scoring and CRITICAL issue caps.

Structured communication evaluation framework used by the brutal-critic-text agent. Defines 6 evaluation categories (clarity of purpose, tone & professionalism, brevity, call to action, audience awareness, formatting & scannability) for emails, Slack messages, proposals, cover letters, and professional messages. Includes report template with mandatory improved rewrites.

Structured technical writing evaluation framework used by the brutal-critic-text agent. Defines 7 evaluation categories (accuracy, structure, clarity, completeness, audience awareness, examples & visual aids, actionability) for RFCs, ADRs, READMEs, architecture docs, design specs, and diagrams. Includes report template with mandatory improved rewrites.

Structured thought and idea evaluation framework used by the brutal-critic-text agent. Defines 6 evaluation categories (logical coherence, evidence & support, completeness, feasibility, originality, counter-arguments considered) for evaluating ideas, arguments, reasoning, strategies, and decision proposals. Includes report template with mandatory strengthened version of the argument.

Structured general writing evaluation framework used by the brutal-critic-text agent. Defines 6 evaluation categories (grammar & mechanics, clarity & conciseness, structure & flow, tone & voice, argumentation & logic, engagement & readability) for articles, blog posts, essays, reports, and general prose. Includes report template with mandatory improved rewrites.