Back to browse
GitHub Repository

Catch your AI coding agent's security slips and turn the correction into a local regression eval. Local-first, deterministic, no LLM judge.

3 starsJavaScript

TreeTrace, git records what changed; this records how you steered

by ZionBoggan·Jun 16, 2026·2 points·0 comments

AI Analysis

●●SolidBig BrainNiche Gem

Turns your agent corrections into regression tests without LLM judges.

Strengths
  • Security regression memory flags auth weaknesses and converts fixes into evals
  • Zero dependencies, local-first with no telemetry or cloud uploads
  • MCP server support for agent integration
Weaknesses
  • Agent evaluation space already crowded with LangSmith, Arize, and others
  • Unclear how deterministic evals handle non-deterministic agent behaviors
Target Audience

AI agent developers, ML engineers building autonomous systems

Similar To

LangSmith · Arize · Braintrust

Similar Projects

Security●●●Banger

Feeding agents with session recording data

Wire-protocol parsing blocks dangerous commands without needing agents on endpoints.

Big BrainSolve My Problem
andriosr
201mo ago