# hiveloom

> An agent is a model plus its harness. hiveloom confines an agent to one job and makes success provable with a declared task, tools, context, loop policy, budgets, guardrails, and deterministic validators.

## Use hiveloom as an agent

- Latest agent-readable docs: https://docs.hiveloom.cloud/llms.txt
- Full docs in one context: https://docs.hiveloom.cloud/llms-full.txt
- Raw Markdown introduction: https://docs.hiveloom.cloud/index.md
- Machine-readable documentation versions: https://docs.hiveloom.cloud/versions.json
- Agent instructions in the repository: https://raw.githubusercontent.com/FrancescoMrn/hiveloom/main/AGENTS.md
- Compact skill: https://raw.githubusercontent.com/FrancescoMrn/hiveloom/main/SKILL.md

The CLI is agent-native: commands support JSON output, mutations validate the complete harness and roll back on failure, and the full contract is available through `hiveloom schema --json`.

## Current release

- Release: 1.2.0 (tag `v1.2.0`, on PyPI as `hiveloom==1.2.0`)
- Release notes: https://docs.hiveloom.cloud/release-notes.md
- Frozen documentation: https://docs.hiveloom.cloud/v1.2.0/llms.txt
- Previous frozen documentation: https://docs.hiveloom.cloud/v1.1.0/llms.txt, https://docs.hiveloom.cloud/v1.0.0/llms.txt
- Demo harnesses, by capability: https://docs.hiveloom.cloud/demos.md
- Install: `uv tool install hiveloom`
- Source: https://github.com/FrancescoMrn/hiveloom