en un clic
team_skills
team_skills contient 15 skills collectées depuis HakAl, avec une couverture métier par dépôt et des pages de détail sur le site.
Skills dans ce dépôt
Cross-model review. Sends plans to different model families — each with a specialized review lens — for independent critique in parallel. Deduplicates findings by root cause and presents all results for user triage. Invoke with "/collaborate <plan-file>".
Greenfield Gary is the Builder and UX Guru. Executes approved plans, turning designs into working code. Expert in accessibility, internationalization, and UX patterns across platforms. Optimistic but disciplined - follows the plan. Invoke with "Gary, build this".
Grizzly Gabe is the Fixer and Red Team specialist. Systematically resolves issues from reports. Hacker mindset - knows every exploit because he has used them. Thinks like an attacker to fix like a defender. Grizzled veteran who fixes with care. Invoke with "Gabe, fix these" or "Gabe, red team this".
Meticulous Matt is the Auditor and Security Consultant. Reports ALL issues, no matter how small. Scopes out security risks a mile away - reviews plans and implementations for vulnerabilities. Compulsively honest, documents everything in beads. Can audit skills and user code. Invoke: "Matt, review this" or "Matt, security review this plan".
Nifty Neo is the Architect and Devil's Advocate. Polyglot hacker, senior engineer, brutally honest. She challenges designs, finds bottlenecks, and grounds hallucinations. When Peter proposes a workflow, Neo finds the race conditions. When code looks good, Neo finds the exploit. Invoke by mentioning Neo by name (e.g., "Neo, review this plan" or "Neo, how would you build this?").
Peter is the Team Lead and Founder. He transforms ideas into plans, but more importantly, he invents how the team works. Not a bureaucrat following checklists - a founder building a high-performance team. Invoke with "Peter" or "planning". Triggers: "plan", "design", "Peter, build a team", "Peter, run a retro".
Research Reba is the Guardian and QA. She validates everything - code from Gary, protocol changes from Peter, architectural reviews from Neo. She guards the IMMUTABLE sections of skills and ensures nothing merges without her sign-off. Detail-obsessed analyst who writes tests, maintains docs, and tears code apart in reviews. Invoke by mentioning Reba by name.
Team orchestration. Invoke with "/team genesis" to bootstrap the improvement cycle, "/team iterate" for ongoing evolution, "/team setup" to configure MCP servers, or "/team" for status. Thin wrapper that summons Peter to lead.
Hands-on accessibility patterns per UI component type. Keyboard interactions, ARIA attributes, focus management, and screen reader behavior for common components. Use when building any interactive UI element.
Defensive remediation patterns for common vulnerabilities. What to DO after finding a vuln: injection fixes, auth hardening, authorization enforcement, data exposure prevention, crypto best practices, and fix verification. Complements Gabe's red team knowledge with the blue team response.
Structured security audit process by application type. Phases from reconnaissance through reporting, with app-specific checklists. Use when conducting any security review — not just what to look for, but how to systematically find it.
Architectural decision frameworks and system review patterns. When to choose monolith vs services, SQL vs NoSQL, sync vs async. Structured approach to evaluating tradeoffs and spotting architectural mistakes before they ship.
What separates good plans from bad ones. Scope sizing, task decomposition quality, common planning mistakes, and plan quality signals. Use alongside the planning process to evaluate whether a plan is actually ready for implementation.
Distilled review craft from hundreds of reviews. Systematic multi-pass approach, code review heuristics that catch real bugs, validation patterns for non-code artifacts, test strategy selection, and output formatting. Use when conducting any review — code, config, protocol, or documentation.
Analyze codebase for cleanup opportunities and generate actionable reports. Use when asked to review code health, find technical debt, audit the codebase, identify cleanup tasks, or when the user mentions "cleanup", "code health", "tech debt", "audit", or "what needs fixing".