Skill Claude CodeCodex
Use when you need to reverse engineer an existing project to understand its business domain, features, workflows, user roles, business rules, and integrations — auto-detects project type (backend, frontend, iOS, Android, or mixed) and produces a report written for business analysts, product owners, and non-technical…
Skill Claude CodeCodex
Use when you need a deep code review of a pull request, branch, or changeset — auto-detects project type, performs structural, semantic, architectural, regression-risk, performance, security, test-alignment and communication-quality analysis of the diff (not the whole repo), and delivers prioritised file:line findings…
Skill Claude CodeCodex
Read .github/skills/cognia-tech/SKILL.md and follow every instruction defined there exactly.
Skill Claude CodeCodex
Use when you need a test engineering analysis of an existing project — auto-detects project type (backend, frontend, iOS, Android, or mixed), audits existing test coverage for functional and structural correctness, identifies missing unit/integration/e2e tests, assesses test quality, and delivers a prioritised list of…
Skill Claude CodeCodex
UI/UX design skill for system redesign. Act as a senior master UI/UX developer. Use when: designing user interfaces for modernized application, creating wireframes mockups design systems, defining user journeys for web React and mobile iOS Android, applying WCAG accessibility standards, building responsive…
Skill Claude CodeCodex
Use when you need a UI/UX analysis of an existing project — page and screen inventory, user flow mapping, design consistency, component library assessment, accessibility audit, responsiveness, interaction patterns, and UX improvement recommendations.
Cursor rule Cursor
Use when you need to record a single architectural decision as an Architecture Decision Record (ADR) through an interactive interview. Act as a principal architect. Use when: recording a new architectural decision, documenting why a technology/approach was chosen, updating an existing ADR, superseding a prior…
Cursor rule Cursor
Use when you need to author a full Application System Design Document (ASDD) through an interactive interview across all 22 sections of docs/asdd-template.md. Act as a principal architect. Use for greenfield ideas, existing-system documentation, target-state design, migration planning, or updates to an existing ASDD.…
Instructions file CodexOpenCode
Instructions for atmosware/cognia-agents, covering cognia — agent roster, agent index, how to use and cross-agent rules.
Agent Claude Code
Use after cognia-android has produced its report. Proposes and (with human approval) applies concrete fixes to Android issues: Kotlin pattern improvements, Gradle dependency updates, manifest corrections, navigation fixes, Room query improvements, coroutine dispatcher fixes, and Play Store readiness issues.
Agent Claude Code
Use after cognia-arch has produced its report. Proposes a target architecture redesign — new component boundaries, service decomposition, data flow, technology recommendations, and a phased migration path — and presents it for human approval before finalising. Never modifies source code; always produces a suggestions…
Agent Claude Code
Use after cognia-backend has produced its report. Proposes and (with human approval) applies concrete fixes to backend issues: endpoint improvements, service refactoring, input validation, auth hardening, database query corrections, API contract improvements, error handling standardisation, and config cleanup.
Agent Claude Code
Use after cognia-frontend has produced its report. Proposes and (with human approval) applies concrete fixes to frontend issues: component improvements, lazy loading, state management, routing guards, form validation, bundle configuration, accessibility attributes, and performance patterns.
Agent Claude Code
Use after cognia-ios has produced its report. Proposes and (with human approval) applies concrete fixes to iOS issues: Swift pattern improvements, navigation refactoring, data persistence corrections, dependency updates, memory management fixes, App Store readiness issues, and missing protocol conformances.
Agent Claude Code
Use after cognia-perf has produced its report. Proposes and (with human approval) applies concrete performance fixes across detected platforms: query optimisation, caching layers, bundle splitting, lazy loading, async parallelism, main-thread unblocking, and image optimisation configuration.
Agent Claude Code
Use after cognia-po has produced its report. Transforms raw product findings into a refined, actionable product backlog — well-formed user stories with acceptance criteria, prioritised epics, gap-filling feature specifications, and a roadmap recommendation — presented for human approval before finalising. Never…
Agent Claude Code
Use after cognia-reverse has produced its report. Transforms reverse-engineering findings into formal, structured documentation — requirements specifications, domain model documents, process flow documents, and gap analysis — presented for human approval before finalising. Never modifies source code; always produces a…
Agent Claude Code
Use after cognia-review has produced its report. Proposes and (with human approval) applies concrete fixes to the changeset under review: addressing Blocker/Major/Minor findings in the diff, regenerating affected tests, correcting PR metadata (title/description/commit messages), and producing a re-review checklist.…
Agent Claude Code
Use after cognia-sec has produced its report. Proposes and (with human approval) applies concrete security fixes across detected platforms: parameterised queries, auth hardening, security headers, CORS restriction, secret externalisation, input validation, session configuration, and mobile data storage corrections.…
Agent Claude Code
Use after cognia-tech has produced its report. Proposes and (with human approval) applies concrete fixes to technical quality issues: dead code removal, dependency updates, linting config, anti-pattern corrections, test coverage scaffolding, and code structure improvements — across any detected platform.
Agent Claude Code
Use after cognia-test has produced its report. Proposes and (with human approval) writes the missing tests identified in the report: unit tests, integration tests, and e2e test stubs — across any detected platform. Each test file or test case requires explicit approval before it is written.
Agent Claude Code
Use after cognia-ux has produced its report. Proposes and (with human approval) applies concrete UX fixes: accessibility attributes, focus management, semantic HTML, copy corrections, and CSS/layout improvements. For changes requiring visual design decisions, produces detailed design specifications as suggestions.
Skill Claude CodeCodex
Use after cognia-android has produced its report. Proposes and (with human approval) applies concrete fixes to Android issues: Kotlin pattern improvements, Gradle dependency updates, manifest corrections, navigation fixes, Room query improvements, coroutine dispatcher fixes, and Play Store readiness issues.
Skill Claude CodeCodex
Use after cognia-arch has produced its report. Proposes a target architecture redesign — new component boundaries, service decomposition, data flow, technology recommendations, and a phased migration path — and presents it for human approval before finalising. Never modifies source code; always produces a suggestions…