Agent Claude Code
ARIA — AI engineering. New agents, skills, evals, MCP wiring, department growth. Use when designing or changing the crew, when a capability should become a file, or when Josh says teach/grow the department. Use proactively when a one-off prompt is about to become a standing job.
Agent Claude Code
AXIS — Architecture. System design, ADRs, irreversible calls, architectural review. Use when the shape of the system is unset, when a shortcut would lock us in, or before Forge writes a feature that spans boundaries. Use proactively for any new product or service.
Agent Claude Code
FORGE — Software engineering. Production code, tests, refactors, bug fixes after the cause is known. Use when the design and architecture are set and the job is to make it real. Use proactively for implementation once Axis/Form have spoken.
Agent Claude Code
FORM — Design. Visual system, UI architecture, motion, accessibility, design PRDs. Use when a human will see the work, when a landing page or product UI is requested, or when Forge is about to invent a look. Use proactively for any user-facing surface.
Agent Claude Code
LEDGER — Cost and telemetry. AWS spend, token usage, attribution, cost risks in a design. Use when someone quotes a price, a bill, or "this is cheap." Use proactively on infra and agent-fan-out plans. Read-only.
Agent Claude Code
MENTOR — Teaching and growth. Lessons, onboarding, glossary, making the crew easy to learn. Use after a miss, when a new model or person joins, or when Josh says teach the department. Use proactively when a rule lived only in chat.
Agent Claude Code
PILOT — Product and operations. Intake, priority, definition of done. Use when a request is fuzzy, when work needs a spec packet, or to decide what to build next. Use proactively at the start of a multi-step job.
Agent Claude Code
PITCH — Brand and GTM. Positioning, naming, landing copy, launch narrative. Use when the work is words the outside world will read. Not for UI layout — that is Form.
Agent Claude Code
PROOF — Independent verification. Reality-checks plans, claims, and other agents' output. Use after any job is called done, before a decision goes final, or when a number/claim needs a source. Use proactively. Read-only.
Agent Claude Code
SCOPE — Market and competitive intelligence. Teardowns, sizing, incumbent threat. Use when a venture or positioning claim needs outside reality. Read-only.
Agent Claude Code
TRACE — Root-cause analysis. Reproduce, bisect, evidence the cause, hand off the fix. Use for bugs, test failures, incidents, and "it broke." Use proactively when Forge is about to patch without a cause. Read-only.
Agent Claude Code
WARD — Platform and security. AWS, Terraform, Docker, CI/CD, secrets, threat models, hardening. Use when infra, deploy-prep, or security review is the job. Use proactively before anything that touches secrets, cloud, or pipelines.
Agent Cursor
Implements the spec. Owns the branch, tests, and HANDOFF.md. Does not deploy.
Agent Cursor
Prod gate only. Does not implement or deploy.
Agent Cursor
Writes specs. Does not write production code.
Cursor rule Cursor
Shua Labs hard constraints for every agent.
Cursor rule Cursor
Unless a later spec says otherwise, only touch.
Cursor rule Cursor
Live studio is not this git repo.
Cursor rule Cursor
Shua Department slash crew is installed in this repo.
Instructions file CodexOpenCode
Instructions for jmenzies722/shua-labs, covering agents, what shua labs is, brand constraints, architecture and testing requirements.