Rashomon keeps an independent record of what your Claude Code agent actually did — every tool call, subagent and failed command — and puts it beside the agent own summary, flagging discrepancies at the end of the turn in one line. It is local-only with no telemetry, and the full report includes what subagents did, which the main conversation does not show. Claude Code only, on macOS and Linux, at alpha.
Be the first to rate Rashomon. Your feedback helps other users decide.
The end-of-turn line flags when the agent claims all tests pass but something failed.
The full report lists subagent activity the main conversation hides.
The record stays on your machine with no telemetry.
Similar tools in the same category — compare side-by-side
Rowan is a static security scanner for AI applications. It reads source code and model files and rep...
Laya guard is a free local guardrail for AI coding agents. Every tool call is judged on-device befor...
BitRun is a zero-compile runner and terminal dashboard for 1-bit and 1.58-bit ternary language model...
Guides and comparisons about code tools
Four Show HN launches that verify agent work rather than trusting its summary: an execution record, an AI-app scanner, a local guardrail and adversarial tests.
A 2026 guide to AI code tools for data science: GitHub Copilot, Codeium, CodeWhisperer, Cursor, Codium AI, Replit, Noteable, Hugging Face, LangChain, Phind,
Three shifts have changed what AI does for developers, from autocomplete to repo-wide edits to review agents. What each category is actually good for.
An honest 2026 look at AI bug-fixing tools: Codium AI (Qodo), BugHunter AI, DeepCode, GitHub Copilot, and why a human must verify the fix.
A 2026 comparison of AI API-testing tools: AutoTest AI, TestGuard AI, APIBuilder AI, Bito. Covers E2E tests, auto-docs, and why human review still matters.