# RQM Agent Ingestion Guide (Full)

## Platform identity

RQM is a quantum software and API platform with a conservative, API-first posture. It is designed for circuit validation, optimization, execution routing, quaternionic/SU(2) math utilities, geometry-native analysis, and RQM Studio workflows.

Canonical docs URL: https://docs.rqmtechnologies.com/

## Public contract boundaries

Treat public contracts as the implementation boundary:

- Public circuit intake is schema-driven (`rqm-circuits`, documented schema target).
- Public behavior is defined by documented API routes and schemas (API docs + Swagger).
- Response envelope structure and documented fields are contract material.

Do **not** treat internal optimizer representations (including `u1q`) as public API contracts.

## Production vs research boundary

Production/documented scope:

- Documented API routes and schemas.
- Validation/analyze/optimize/execution behavior described in product and API docs.
- Verification/trust posture represented in shipped docs.

Research scope:

- Speculative theory framing and exploratory workflow ideas.
- Proposed or unshipped integrations/tooling.
- Performance or physical-advantage claims without documented benchmark evidence.

Rule: label research explicitly as research and keep production claims bounded.

## Known repo ecosystem

Key components referenced across docs:

- `rqm-core`: shared math and analysis primitives.
- `rqm-circuits`: public circuit payload boundary.
- `rqm-compiler`: internal optimization engine.
- `rqm-api`: service/API boundary.
- Execution bridges: `rqm-qiskit`, `rqm-braket`, `rqm-pennylane`.
- `rqm-studio`: workflow UX/orchestration layer over the API.

Treat this as architecture context, not permission to infer undocumented external contracts.

## API-first usage expectations

For implementation tasks:

1. Read `/llms.txt`.
2. Read `/agents/` entry pages and relevant task recipe.
3. Read `/agent-contract.json` and compliance pages.
4. Confirm exact route/schema behavior in `/api/overview/` and Swagger.
5. Build validate-first workflows (`validate -> analyze/optimize -> execution`).
6. Save reproducible artifacts for each behavior claim.

## What not to infer

- Do not invent endpoints, methods, request fields, or response fields.
- Do not infer internal pass ordering or optimizer IR semantics as public behavior.
- Do not infer provider capabilities beyond documented execution bridge behavior.
- Do not convert research language into production guarantees.

If docs are unclear, report the uncertainty instead of guessing.

## Citation and evidence expectations

When producing outputs (PRs, design notes, implementation plans, audits):

- Cite the docs pages used for routes, schemas, and boundary rules.
- Tie each behavior claim to a documented source or reproducible artifact.
- Distinguish “documented production behavior” from “research interpretation.”

## Suggested artifact outputs

Prefer outputs that another engineer or agent can replay:

- Request payload JSON and response JSON.
- Validation/optimization before-vs-after diffs.
- Route/schema confirmation notes (from API docs/Swagger).
- Workflow step logs for Studio orchestration.
- Contract compliance checklist results for PR audits.

## High-value links

- Agent Portal: /agents/
- Prompt Pack: /agents/prompts/
- Agent Contract: /agents/contract/
- Contract Checklist: /agents/contract-checklist/
- Contract Tests: /agents/contract-tests/
- Agent Recipes: /agents/recipes/
- API Overview: /api/overview/
- Swagger UI: https://rqm-api.onrender.com/docs