# lavra

> A plugin with compound engineering workflows and memory for AI coding agents

- **URL**: https://www.freshcrate.ai/projects/lavra
- **Author**: roberto-mello
- **Category**: AI Agents
- **Latest version**: `v0.7.6` (2026-04-21)
- **License**: MIT
- **Source**: https://github.com/roberto-mello/lavra
- **Homepage**: https://lavra.dev
- **Language**: Shell
- **GitHub**: 37 stars, 1 forks
- **Registry**: github
- **Tags**: `agents`, `ai`, `beads`, `claude-code`, `coding`, `cortex-code`, `gemini-cli`, `multi-agent`, `shell`

## Description

A plugin with compound engineering workflows and memory for AI coding agents

## Recent releases

| Version | Date | Urgency | Changes |
| --- | --- | --- | --- |
| `v0.7.6` | 2026-04-21 | High | **Full Changelog**: https://github.com/roberto-mello/lavra/compare/v0.7.5...v0.7.6 |
| `v0.7.5` | 2026-04-19 | High | ## What's Changed * lavra-ceh: split lavra-work + add Karpathy principles by @roberto-mello in https://github.com/roberto-mello/lavra/pull/9 * lavra-ziy: caveman-lite prose guide, rg check, agent compression by @roberto-mello in https://github.com/roberto-mello/lavra/pull/10 * lavra-3hp.2: scope lavra-review to introduced diff, split pre-existing findings by @roberto-mello in https://github.com/roberto-mello/lavra/pull/12   **Full Changelog**: https://github.com/roberto-mello/lavra/compare/v0.7. |
| `v0.7.4` | 2026-04-08 | High | **Full Changelog**: https://github.com/roberto-mello/lavra/compare/v0.7.3...v0.7.4 |
| `v0.7.3` | 2026-04-08 | Medium | **Full Changelog**: https://github.com/roberto-mello/lavra/compare/v0.7.2...v0.7.3 |
| `v0.7.2` | 2026-04-08 | Medium | ## What's Changed * fix: arithmetic increment crashes installer steps 5-7 under set -e by @xanharris in https://github.com/roberto-mello/lavra/pull/6 * fix(memory-capture): scope extraction to first matching bd comment line by @xanharris in https://github.com/roberto-mello/lavra/pull/7 * v0.7.2: review_scope config, configuration page, agent colors, memory hardening by @roberto-mello in https://github.com/roberto-mello/lavra/pull/8  ## New Contributors * @xanharris made their first contribution |
| `v0.7.1` | 2026-03-27 | Medium | ## What's Changed * feat: .lavra/ migration, ASCII banner, npm publish as @lavralabs/lavra by @roberto-mello in https://github.com/roberto-mello/lavra/pull/3   **Full Changelog**: https://github.com/roberto-mello/lavra/compare/v0.7.0...v0.7.1 |
| `v0.7.0` | 2026-03-18 | Low | ## What's Changed * bdcompound-9f5 epic work by @roberto-mello in https://github.com/roberto-mello/lavra/pull/2   **Full Changelog**: https://github.com/roberto-mello/lavra/compare/v0.6.8...v0.7.0 |
| `v0.6.8` | 2026-03-11 | Low | **Full Changelog**: https://github.com/roberto-mello/beads-compound-plugin/compare/v0.6.7...v0.6.8 |
| `v0.6.7` | 2026-03-05 | Low | **Full Changelog**: https://github.com/roberto-mello/beads-compound-plugin/compare/v0.6.6...v0.6.7 |
| `v0.6.6` | 2026-02-23 | Low | ## What's Changed  ### Improved: Memory recall is now a required gate, not a suggestion  Agents were acknowledged to be skipping memory recall before starting work — even when the instructions were explicit. This release makes recall structurally harder to skip across both `beads-work` and `beads-parallel`.  **`/beads-work`** - Phase 1 step 2 now marked as "required — do not skip" - Agent must output recall results (or "No relevant knowledge found") before proceeding to implementation  **`/beads |

## Citation

- HTML: https://www.freshcrate.ai/projects/lavra
- Markdown: https://www.freshcrate.ai/projects/lavra.md
- Dependencies JSON: https://www.freshcrate.ai/api/projects/lavra/deps

_Generated by freshcrate.ai. Indexes github releases for AI-agent ecosystem packages._