MCP Tools Directory
← Browse integrations

Developer tools

culprit

Free account · sign-in required

Contact culprit

Loading secure sign-in…

Tools for autonomous AI agents

About this integration

Deterministic, offline root-cause analysis and pre-commit fix verification for coding agents.

Transport
stdio
Authentication
none
Initial setup
none
Runtime
unattended
Evidence
documented
Version
0.4.0
Package
culprit
Last compatibility test
Not independently tested

Connect your agent

culprit

Publisher README and Python manifest reviewed from pinned captures. Package was not executed and benchmark results were not independently reproduced.

Capabilities: Rank commits that may have introduced a bug, Check whether a proposed fix covers affected call sites, Estimate change risk and blast radius, Select impacted tests and reconstruct line history, Analyze a stack trace without a diff

Connected profiles

Additional details

Registry identifier

io.github.noordeen123/culprit

Source ↗ · Checked 2026-09-17
Published version

0.4.0

Source ↗ · Checked 2026-09-17
Metadata license

CC0-1.0; package licenses are separate

Source ↗ · Checked 2026-09-17
Package ecosystem

pypi

Source ↗ · Checked 2026-09-17
Source artifact

culprit

Source ↗ · Checked 2026-09-19
Source manifest version

unspecified

Source ↗ · Checked 2026-09-19
Unattended configuration

Publisher documents an offline, read-only stdio MCP server launched through uvx and returning deterministic structured analysis without running tests or a model by default.

Source ↗ · Checked 2026-09-19
Authentication and prerequisites

The MCP server is documented as offline with no API key and no network requirement; optional CLI integrations outside the default MCP path may use repository-provider tokens.

Source ↗ · Checked 2026-09-19
Connection configuration

The documented MCP connection launches uvx --from culprit[mcp] culprit-mcp against the current local repository.

Source ↗ · Checked 2026-09-19
Review scope

Publisher documentation reviewed; package and integration endpoint not executed or independently security-audited.

Source ↗ · Checked 2026-09-19