Struere Developer

Struere is an AI-native full-stack agent platform. Define agents, data types, roles, triggers, routers, and tools as TypeScript — then sync to dev and deploy to prod.

Rules

1. Ensure struere dev is running. Before making any changes, verify struere dev is running in the background. If it's not running or not available, use bunx struere sync after each change. Never ask the user whether to sync — just keep changes synced.
2. Fetch docs before writing code. NEVER write or modify Struere code without first fetching the relevant URL from the documentation table below. Every time. No exceptions. Do not work from memory.
3. Use struere chat to test. After creating or modifying an agent, test it with struere chat <slug> to simulate conversations before writing evals. Run multiple test conversations to discover edge cases.
4. Ask before assuming. Before creating an agent, ask what it should do, who it serves, and what tools it needs. Do not guess requirements.
5. One export per file. Each file in agents/, entity-types/, roles/, triggers/, routers/ exports a single default. Only tools/index.ts exports all custom tools.
6. Slugs are identities. Renaming a slug creates a new resource — it does not rename the existing one.
7. Environments are isolated. struere dev syncs to development + eval (eval excludes triggers and routers). struere deploy syncs to production.
8. Roles work for both users and agents. A role's policies/scope rules/field masks apply to whoever is assigned to it. Users get roles via the dashboard; agents declare them in code via defineAgent({ roles: [...] }). Don't conflate agentAccess (dashboard ACL) with agent.roles (permission inheritance). Roles use slug for identity (lowercase, alphanumeric + hyphens); name is a display label. Always declare slug on new roles — it'll be required in a future major version.
9. Keep tools under 5 per agent. Split into specialist agents with agent.chat if you need more.
10. Use bun, never npm. Always bun install, bun run, bunx.
11. Use struere keys for credentials. Get API keys via struere keys create from the CLI — don't direct users to the dashboard for routine key management.
12. Keep CLI updated. Run bun update struere to ensure you're on the latest SDK/CLI version before starting work.

Documentation

Core SDK (fetch before defining any resource):

Task	URL
Define an agent	https://docs.struere.dev/sdk/define-agent.md
Define a data type	https://docs.struere.dev/sdk/define-data.md
Define a role / permissions	https://docs.struere.dev/sdk/define-role.md
Define a trigger	https://docs.struere.dev/sdk/define-trigger.md
Define a router	https://docs.struere.dev/sdk/define-router.md
Define custom tools	https://docs.struere.dev/sdk/define-tools.md
Full SDK reference	https://docs.struere.dev/llms-sdk.txt

Tools & Prompts (fetch when configuring agent behavior):

Task	URL
System prompt templates	https://docs.struere.dev/tools/system-prompt-templates.md
Built-in tools list	https://docs.struere.dev/tools/built-in-tools.md
Custom tools guide	https://docs.struere.dev/tools/custom-tools.md
Model configuration	https://docs.struere.dev/reference/model-configuration.md

Integrations (fetch when adding channels):

Task	URL
WhatsApp	https://docs.struere.dev/integrations/whatsapp.md
Google Calendar	https://docs.struere.dev/integrations/google-calendar.md
Airtable	https://docs.struere.dev/integrations/airtable.md
Email (Resend)	https://docs.struere.dev/integrations/resend.md
Flow Payments	https://docs.struere.dev/integrations/flow-payments.md
Voice (Twilio)	https://docs.struere.dev/integrations/voice.md
Embeddable widget	https://docs.struere.dev/integrations/embeddable-widget.md
All integrations	https://docs.struere.dev/llms-integrations.txt

API & Reference (fetch when integrating externally):

Task	URL
Chat API	https://docs.struere.dev/llms-api.txt
OpenAPI spec	https://docs.struere.dev/openapi.yaml
JavaScript client	https://docs.struere.dev/api/javascript-client.md
API keys CLI	https://docs.struere.dev/cli/keys.md
Error codes	https://docs.struere.dev/reference/error-codes.md
Rate limits	https://docs.struere.dev/reference/rate-limiting.md
Project structure	https://docs.struere.dev/reference/project-structure.md

Knowledge Base (fetch when debugging or architecting):

Task	URL
RBAC setup	https://docs.struere.dev/knowledge-base/how-to-set-up-rbac.md
Debug permission denied	https://docs.struere.dev/knowledge-base/how-to-debug-permission-denied.md
Handle tool errors	https://docs.struere.dev/knowledge-base/how-to-handle-tool-errors.md
Manage environments	https://docs.struere.dev/knowledge-base/how-to-manage-environments.md
Test with evals	https://docs.struere.dev/knowledge-base/how-to-test-with-evals.md
Multiple agents	https://docs.struere.dev/knowledge-base/how-to-use-multiple-agents.md
Template variables	https://docs.struere.dev/knowledge-base/how-to-use-template-variables.md
Add chatbot to website	https://docs.struere.dev/knowledge-base/how-to-add-chatbot-to-website.md
Best practices	https://docs.struere.dev/knowledge-base/best-practices.md
Debug triggers	https://docs.struere.dev/platform/triggers.md

CLI & Full Docs:

Task	URL
CLI commands	https://docs.struere.dev/llms-cli.txt
Platform concepts	https://docs.struere.dev/llms-platform.txt
Full documentation	https://docs.struere.dev/llms-full.txt
Docs index	https://docs.struere.dev/llms.txt

What are you trying to do?

• Building a new project from scratch → see workflows/build-new-project.md
• Adding features to an existing project → see workflows/add-to-project.md
• Calling Struere from an external app (web/mobile/server) → see patterns/external-integration.md
• Debugging or fixing issues → see workflows/debug-and-fix.md
• Making an architecture decision (model, tools, multi-agent) → see decision-frameworks.md
• Configuring per-agent permissions (different agents see different data) → see patterns/agent-complex.md
• Need a code example → see patterns/ directory
• Writing or reviewing code → check platform-gotchas.md before submitting
• Designing agents, prompts, tools, or evals → see best-practices.md