# AGENTS.md — Build guide for Admix

## Project scope
Send the same reviewed prompt to two selected models, stream independent response columns, and save a comparison with editable notes and Markdown export.

Catalogue verdict: kinda. A local chat UI with side-by-side replies is a weekend project when it calls OpenRouter, but Admix's value depends on access to many hosted models, a current catalog, provider routing, and a maintained account and billing layer. You can reproduce the comparison workflow. You cannot reproduce the model access and service layer for a flat $10 subscription.
Use the implementation prompt below to define the deliverable. Complete each phase's acceptance checks before extending the scope.

## Working agreement
- Inspect the repository and its existing instructions before choosing paths, dependencies or commands. Keep one coherent stack and explain changes to the proposed architecture.
- Plan a vertical slice that accepts a real input and produces the useful output described below. Persist only the state the prompt calls for; respect memory-only and upstream-managed workflows. Use fixtures only when they are clearly labelled.
- After scaffolding, document the actual install, development, check and build commands in README and keep them synchronized with the package or project manifest. Do not report commands as successful unless they ran.
- Work in small steps. At handoff, list implemented flows, checks actually performed, remaining blockers, and any credentials or provider setup the owner must supply.
- Do not publish, spend money, contact customers, delete source data or run irreversible migrations without the project owner's authorization.

## Prerequisites
- Runtime and tools: TypeScript, Node, SQLite and a React review screen with one configurable model adapter.
- Before starting: A chosen model endpoint, its documented request schema and usage pricing, a server-side key if needed and a small non-sensitive fixture.

## Stack and architecture
- TypeScript, Node, SQLite and a React review screen with one configurable model adapter
- Data design: Store Comparison, PromptRevision, ModelAttempt and ResponseChunk; each column has its own ordered chunks, status and usage record. Regeneration creates a new attempt.
- Setup: A chosen model endpoint, its documented request schema and usage pricing, a server-side key if needed and a small non-sensitive fixture

## Security and data integrity
- Keep source evidence, model/config version, draft output and reviewer changes separately. Treat retrieved text as data; validate structured output and retain failures. Never silently send private material to a fallback provider.
- Load supported model identifiers from the chosen gateway and retain a dated catalogue snapshot. Display estimated costs separately from reported usage; sanitized Markdown does not make model instructions trustworthy.
- Keep secrets outside client bundles and exported projects; document what leaves the device and make retention/deletion controls visible.

## Agent implementation rules
- Project rule — domain: Store Comparison, PromptRevision, ModelAttempt and ResponseChunk; each column has its own ordered chunks, status and usage record. Regeneration creates a new attempt.
- Project rule — scope and recovery: Load supported model identifiers from the chosen gateway and retain a dated catalogue snapshot. Display estimated costs separately from reported usage; sanitized Markdown does not make model instructions trustworthy.
- Project rule — acceptance: Make one model rate-limit while the other completes; the successful column remains readable and retrying the failed column does not resend its sibling.
- Project rule — delivery: document real setup commands and permissions; do not claim a build, accuracy level, performance result or security certification that has not been demonstrated.

## Optional agent skills and references
- Recommended skill: [web-design-guidelines](https://github.com/vercel-labs/agent-skills/blob/main/skills/web-design-guidelines/SKILL.md) — review keyboard access, focus, validation, error recovery and the readable work/review interface or HTML report. Follow the maintainer's installation instructions and match its requirements to the chosen runtime.
- Recommended skill: [sharp-edges](https://github.com/trailofbits/skills/blob/main/plugins/sharp-edges/skills/sharp-edges/SKILL.md) — review configuration and API defaults against the app-specific invariants and recovery boundaries above; this is not a security certification. Follow the maintainer's installation instructions and match its requirements to the chosen runtime.

Read the linked SKILL.md and its dependencies before adding a skill. Select only the skills matching this project's runtime and task; their documentation does not supply API access, credentials or approval to perform external actions. Pin the reviewed revision where the tool supports it. Follow the chosen agent's documented project-level installation mechanism.

## Distribution ideas
These are optional planning notes. Obtain the owner's approval before publishing or contacting anyone.
- Demonstrate this working slice using synthetic or explicitly authorized non-sensitive examples: Send the same reviewed prompt to two selected models, stream independent response columns, and save a comparison with editable notes and Markdown export.
- Share a synthetic example export and the acceptance walkthrough; keep real customer, health, financial and source data private: Make one model rate-limit while the other completes; the successful column remains readable and retrying the failed column does not resend its sibling.
- State the limits before asking someone to replace their existing tool: Load supported model identifiers from the chosen gateway and retain a dated catalogue snapshot. Display estimated costs separately from reported usage; sanitized Markdown does not make model instructions trustworthy.

## Engineering roadmap
1. Phase 1 — Pin the working slice and create its example input: Send the same reviewed prompt to two selected models, stream independent response columns, and save a comparison with editable notes and Markdown export. Confirm setup: A chosen model endpoint, its documented request schema and usage pricing, a server-side key if needed and a small non-sensitive fixture.
2. Phase 2 — Implement persistence and write-time invariants before decorating the UI: Store Comparison, PromptRevision, ModelAttempt and ResponseChunk; each column has its own ordered chunks, status and usage record. Regeneration creates a new attempt.
3. Phase 3 — Connect the working view to real saved state. Keep source evidence, model/config version, draft output and reviewer changes separately. Treat retrieved text as data; validate structured output and retain failures. Never silently send private material to a fallback provider.
4. Phase 4 — Expose the app-specific limits and recovery path in context: Load supported model identifiers from the chosen gateway and retain a dated catalogue snapshot. Display estimated costs separately from reported usage; sanitized Markdown does not make model instructions trustworthy.
5. Phase 5 — Walk through this concrete acceptance case and preserve its exported evidence: Make one model rate-limit while the other completes; the successful column remains readable and retrying the failed column does not resend its sibling. Finish the README and backup/restore instructions; report unfinished capabilities explicitly.

## Paid-product capabilities outside this build
- Admix's bundled access to many providers under one subscription
- A maintained model catalog and provider routing
- Cross-device sync and account history
- Built-in image, search, and document workflows
- Usage limits, billing, and support handled for you

## Implementation prompt
Build the following focused alternative to Admix. This is a deliberately limited personal or small-team substitute, not parity with the paid service.

WORKING SLICE
Send the same reviewed prompt to two selected models, stream independent response columns, and save a comparison with editable notes and Markdown export.

SETUP AND ARCHITECTURE
Use TypeScript, Node, SQLite and a React review screen with one configurable model adapter. Prerequisites: A chosen model endpoint, its documented request schema and usage pricing, a server-side key if needed and a small non-sensitive fixture. Before integrating anything, record actual versions and permissions, plus model files or provider limits only where used, in the README; make unavailable dependencies visible rather than simulating success.

DOMAIN MODEL AND INVARIANTS
Store Comparison, PromptRevision, ModelAttempt and ResponseChunk; each column has its own ordered chunks, status and usage record. Regeneration creates a new attempt.

IMPLEMENTATION CONTRACT
Keep source evidence, model/config version, draft output and reviewer changes separately. Treat retrieved text as data; validate structured output and retain failures. Never silently send private material to a fallback provider. Provide an input/setup view, the main work view, and a review/export view appropriate to this workflow. Preserve the last saved state if a job or save fails. Include empty, loading, permission-denied, partial and retryable-error states. Log identifiers and error categories without secret values or unnecessary private content.

APP-SPECIFIC BOUNDARY AND RECOVERY
Load supported model identifiers from the chosen gateway and retain a dated catalogue snapshot. Display estimated costs separately from reported usage; sanitized Markdown does not make model instructions trustworthy.

ACCEPTANCE SCENARIO
Make one model rate-limit while the other completes; the successful column remains readable and retrying the failed column does not resend its sibling. Also reopen the app after an interrupted operation, confirm the saved record/export remains inspectable, and document the recovery action. These are implementation acceptance requirements, not a claim that this guide has been tested.

DELIVERY
Deliver a runnable repository with migrations or project-format versioning, a non-sensitive example, environment/permission setup, the exact manual acceptance steps, and a backup/export-and-restore walkthrough. Implement the working slice before optional integrations; list any deferred paid-product capabilities honestly. Do not add capabilities outside the working slice just to resemble the original product.

PROJECT RULES FOR AGENTS.md
Keep the domain invariants above executable at the write boundary. Propose scope changes before adding providers or permissions. Never fabricate source evidence, publish results, identity matches or successful delivery. Preserve user originals and require an explicit confirmation for destructive changes or external publication.

## Completion evidence
Demonstrate the prompt's acceptance scenarios against the scoped workflow. Include setup from a clean checkout and failure recovery. Check persistence across restart and export/restore only for the state the prompt says to store; for memory-only tools, confirm that temporary content is discarded as specified. Record actual results and remaining limitations. A detailed plan alone does not establish a working replacement.