buildlog 0.8.0

Engineering notebook for AI-assisted development

buildlog

The Only Agent Learning System You Can Prove Works

Falsifiable claims. Measurable outcomes. No vibes.

RE: The art — Yes, it's AI-generated. Yes, that's hypocritical for a project about rigor over vibes. Looking for an actual artist to pay for a real logo. If you know someone good, open an issue or DM me. Budget exists.

Read the full documentation

---

Everyone's building "agent memory." Blog posts announce breakthroughs. Products ship with "learning" in the tagline. Ask them one question: How do you know it works?

buildlog gives you the infrastructure to answer with data. It captures engineering knowledge from work sessions, extracts rules, selects which rules to surface using a Thompson Sampling bandit, and measures impact via Repeated Mistake Rate (RMR) across tracked experiments.

Features

• Structured capture — Document work sessions as entries with mistakes, decisions, and outcomes
• Rule extraction — Distill and deduplicate patterns into actionable rules
• Thompson Sampling bandit — Automatic rule selection that balances exploration and exploitation
• Experiment tracking — Sessions, mistakes, RMR calculation with statistical rigor
• Review gauntlet — Curated reviewer personas (Security Karen, Test Terrorist) with HITL checkpoints
• Multi-agent support — Render rules to Claude Code, Cursor, GitHub Copilot, Windsurf, Continue.dev
• MCP server — Full Claude Code integration via buildlog-mcp

Quick Start

uv pip install buildlog   # or: pip install buildlog (inside a venv)
buildlog init
buildlog new my-feature
buildlog distill && buildlog skills
buildlog experiment start
# ... work ...
buildlog experiment end
buildlog experiment report

Documentation

Section	Description
Installation	Setup, extras, and initialization
Quick Start	Full pipeline walkthrough
Core Concepts	The problem, the claim, and the metric
CLI Reference	Every command documented
MCP Integration	Claude Code setup and available tools
Experiments	Running and measuring experiments
Review Gauntlet	Reviewer personas and the gauntlet loop
Multi-Agent Setup	Render rules to any AI coding agent
Theory	The math behind Thompson Sampling
Philosophy	Principles and honest limitations

Contributing

git clone https://github.com/Peleke/buildlog-template
cd buildlog-template
uv venv && source .venv/bin/activate
uv pip install -e ".[dev]"
pytest

We're especially interested in better context representations, credit assignment approaches, statistical methodology improvements, and real-world experiment results (positive or negative).

License

MIT License — see LICENSE

---

"Agent learning" without measurement is just prompt engineering with extra steps.

buildlog is measurement.