# YOLOZU

> Use YOLOZU when you need to validate existing vision predictions, evaluate them against labels, or inspect bounded local MCP capabilities.

A compact, optional documentation index for agents. It grants no permissions and does not guarantee search visibility, ranking, or tool selection.

## Start here

- [Agent guide](https://www.toppymicros.com/yolozu/docs/agents.md): When to use YOLOZU, limits, local MCP setup, and starter prompts.

- [Evaluation quickstart](https://www.toppymicros.com/yolozu/docs/start.md): Isolated install, generated inputs, strict checks, and real COCOeval.

- [Capabilities JSON](https://www.toppymicros.com/yolozu/docs/capabilities.json): Source version, exact MCP surface sets, source links, and hashes.

## Reference

- [Installation and dependencies](https://raw.githubusercontent.com/ToppyMicroServices/YOLOZU/main/docs/install.md): Choose a package extra and distinguish installed commands from repository scripts.

- [CLI tool manifest](https://raw.githubusercontent.com/ToppyMicroServices/YOLOZU/main/tools/manifest.json): Command inputs, examples, side effects, dependencies, and maturity.

- [MCP and Actions reference JSON](https://raw.githubusercontent.com/ToppyMicroServices/YOLOZU/main/docs/generated/mcp_actions_tool_reference.json): Exact generated tool surfaces and MCP input schemas; registration is not execution qualification.

- [Predictions validation result schema](https://raw.githubusercontent.com/ToppyMicroServices/YOLOZU/main/docs/schemas/predictions_validation_result.schema.json): Machine-readable strict validation results, errors, and repair disclosure.

- [Stable Python API](https://raw.githubusercontent.com/ToppyMicroServices/YOLOZU/main/docs/python_api.md): Typed input, result, and error interfaces for local validation and evaluation.

- [LLM and MCP integration](https://raw.githubusercontent.com/ToppyMicroServices/YOLOZU/main/docs/llm_integrations.md): Client connection options and workspace boundaries.

- [Bounded image service](https://raw.githubusercontent.com/ToppyMicroServices/YOLOZU/main/docs/image_service_mcp.md): Capabilities, opaque assets and jobs, limits, and abstention semantics.

- [Local OpenAI plugin](https://raw.githubusercontent.com/ToppyMicroServices/YOLOZU/main/docs/openai_image_service_plugin.md): Prepare a local plugin from a reviewed checkout; not a public app listing.

- [Readiness and maturity](https://raw.githubusercontent.com/ToppyMicroServices/YOLOZU/main/docs/production_readiness.md): Keep Stable, Bridge, Benchmark, and Research claims separate.

- [Discovery publishing and measurement](https://raw.githubusercontent.com/ToppyMicroServices/YOLOZU/main/docs/llm_discovery.md): Publish the bundle, verify access, and measure successful use without inferring adoption from crawler traffic.

## Boundaries

Do not infer model quality from a synthetic proof or a successful dry run. Real COCO metrics require the coco extra and labeled inputs.

Do not treat registered MCP tools as guaranteed execution. Runtime, dependency, data, and permission requirements still apply.

The packaged adaptive image service abstains until CNN license review, quality qualification, and activation are completed separately. It is not a ready-to-run hosted detector.

The local OpenAI plugin is not a public ChatGPT directory listing. Provider authentication, hosting, and attachment handoff are separate deployment work.

docs_version identifies the source checkout, not the installed package or latest PyPI release. Source URLs follow a mutable branch; check source_sha256 and provenance.json for this build. Inspect the installed CLI and MCP schemas before invoking tools.