Documentation Index
Each topic has one owning document. Other files link here instead of repeating the content.
Understand the System
| Document | Covers |
| architecture.md | Principles, modules (current vs. target), execution pipeline, AI integration, planned protocol |
| roadmap.md | Delivery status by phase and next priorities |
| benchmarks/flow_format.md | Benchmark: local AI models writing flows in flow.md vs YAML, and why Kinenix keeps flow.md |
| design/expressions.md | Design of expressions in ${...} and and / or / not in conditions (0.3.0b1) |
| design/list-functions.md | Design of sorting, filtering, totals, and comparisons in ${...} (1.1.0) |
| design/loop-control.md | Design of logic.break and loops with times (1.1.0) |
| design/agent-interface.md | Design of the agent interface: exit codes, --json, tools, error codes, confirmation, flow approval, MCP server, job dispatch; foundation in 1.0.0, the rest in 1.5.0 |
| design/flow-deploy.md | Design of deploying flows to workers: signed packages, publish to the Hub, workers pull and install (proposal) |
| design/threat-model.md | Threat model of core, worker, and Hub: what is protected, trust assumptions, threats with their milestone, review plan and findings |
| design/phase-2.md | Phase 2 design: On Error and Finally, subflow context, run records for AI questions, Hub tables and views, redaction (proposal) |
Build Flows
| Document | Covers |
| flow_markdown_spec.md | flow.md syntax, validation, and the compiled form (kinenix compile) |
| actions_reference.md | Every built-in action with parameters and examples |
| project_bundles.md | Bundle layout, config.json, .env, subflows (flow.call), deployment lifecycle, testing against saved pages |
| batch_flows.md | Designing a flow that loops over many items unattended: output folder per run, per-item errors, empty results, completion messages, test pages for JavaScript sites |
| llms.txt | For AI assistants: how a flow is built and which pages to read, served as https://www.kinenix.com/llms.txt |
Run and Operate
| Document | Covers |
| examples.md | The examples that ship with Kinenix, what each does, and which patterns it shows |
| tutorial.md | Tutorial: build a first flow from requirements.md with an AI assistant, validate, run, and break it on purpose (the first_flow example) |
| quickstart.md | Install, create a project with kinenix init, run it, and connect a worker to the Hub in about ten minutes |
| cli_guide.md | kinenix commands and options |
| hub.md | Running the Hub, environment variables, worker authentication, querying runs (kinenix hub query), user views |
| hub_data.md | Hub data dictionary: tables, views, run reports, and example queries, for people and AI |
| logging.md | Execution log layout and JSON format |
| ../kinenix-worker/README.md | Worker CLI and Raspberry Pi setup |
| ../CHANGELOG.md | Changes and upgrade notes of each release |
| compatibility.md | What stays stable for all of 1.x, what may change, deprecations, and upgrading |
| ../SECURITY.md | Supported versions and how to report a vulnerability privately |
| releasing.md | Publishing kinenix, kinenix-hub, and kinenix-worker to PyPI with Trusted Publishing |
| ../kinenix-studio/README.md | Studio design and layout |
Contribute
| Document | Covers |
| ../AGENTS.md | Entry point for AI assistants: commands, rules index, current vs. target state |
| ../.agents/rules/ | Engineering rules: coding, testing, security, errors, logging, git |
Machine-readable schemas: ../schemas/flow.schema.json, ../schemas/execution_log.schema.json.