Agent Codex
Turn shot-plan.json into one renderable HyperFrames composition (compositions/index.html). Everything stays in the HF ecosystem — HTML is the source of truth; a single paused GSAP timeline carries all motion; the engine seeks it. Category-specific build rules live in categories/ /module.md; this file is the shared…
Agent Codex
Turn a request into a shot-plan.json for a short (3–30s) design-led motion graphic. You run in two parts around the asset-sourcing step: Part 1 (plan) before sourcing, Part 2 (design) after. You do NOT write composition code — that's the Builder. Schema: references/shot-plan-ir.md.
Agent Codex
Perform snapshot QA and one in-place repair pass. Dispatch only when Step 5 lint, check, or snapshot review reports a defect. The orchestrator owns final approval and render; this agent never renders.
Agent Claude Code
Deep analysis of AI/Agent infrastructure — sidecar, Claude Agent SDK, AI SDK, agents, skills, MCP, digital twin, and chat pipeline.
Agent Claude Code
Deep analysis of cognia-next frontend architecture — app router, component tree, hooks, stores, i18n, and UI patterns.
Agent Claude Code
Deep analysis of CI/CD, build system, services (signaling, share), configuration, and monorepo tooling.
Agent Claude Code
Deep analysis of the plugin runtime (WASM), workflow engine (React Flow + TS/Rust), and first-party plugin ecosystem.
Agent Claude Code
Deep analysis of the Rust/Tauri backend — services, automation, OCR, workflows, vector store, MCP server, and CLI.
Agent Claude Code
Deep analysis of testing infrastructure, coverage, E2E suites, and quality assurance patterns.
Agent Claude Code
Read-only i18n review for cognia-next changes. Use proactively after touching any .tsx or i18n/messages file — finds NEW hard-coded user-facing strings (the lint:i18n baseline tolerates 487 legacy findings, so new ones slip through until the count rises), and verifies every newly referenced translation key exists in…
Agent Claude Code
Read-only audit that every outbound LLM / embedding / cloud call path in cognia-next passes through the PII redaction gate. Use proactively when a change adds or reroutes prompts, embeddings, connector auto-replies, or twin/memory distillation — anything that sends locally-derived text to a model. Reports bypasses…
Agent Claude Code
Read-only audit for static-export violations in cognia-next. Use proactively after changes that add routes, server-ish dependencies, or Node built-in imports — the app is a Next.js static export (output "export") consumed by Tauri and Capacitor, so server-only code paths silently break production. Flags issues; does…
Agent Claude Code
Runs citation-backed, current web research through Codex CLI with a read-only sandbox. Use only when delegated by the codex-web-search skill.
Agent Claude Code
Read-only review of src-tauri Rust changes for cognia-next-specific traps — parkinglot guards held across .await, detached tokio tasks that hang cargo test, tuple returns that serialize as JSON arrays, unregistered commands, missing capability/ACL entries, and missing.
Agent Claude Code
Read-only audit of the co-located-test rule for cognia-next. Use proactively before claiming a change "done" or before committing — finds new/changed source files under components/, hooks/, lib/, stores/, src-tauri/src/ that lack the required co-located test, and flags coverage-gate risks. Reports findings; does not…
Agent Claude Code
Read-only "built-but-dormant" detector for cognia-next. Use proactively after implementing any new module, component, command, plugin, or initializer — verifies the new code is actually reachable at runtime (imported, mounted, registered, called at bootstrap), because this repo's most recurrent defect is fully-built…
Agent
Read-only correctness, security, and regression reviewer.
Agent
Fast read-only Cognia codebase explorer.
Agent
Isolated implementation worker for bounded Cognia changes.
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: