Impatient ProgrammingIMP

Tools

Measurement

Delivery metrics, and an honest label on what the agents wrote.

Rule 9 says throughput is measured and labeled honestly. Delivery metrics platforms have measured throughput for years. The label is newer. A few tools now record which lines an agent wrote, and the Linux kernel asks for an Assisted-by trailer on every patch an AI helped with.

  • Start here

    git-ai

    Git extension that records which lines AI wrote, with agent and model, in git notes.

    Rule 9 gets real numbers: the share of agent-written code per commit, tracked through rebases.

    Its founders joined OpenAI's Codex team and the project stays Apache-2.0.

    Any languageWorks with Claude Code, Cursor, Copilot, Codex, Gemini Open source, Apache-2.02.8k starsLatest release Sep 2026Source

  • Agent Trace

    Open spec for recording which code ranges AI or humans wrote.

    A shared format keeps attribution intact when a team changes agent vendor.

    The spec is v0.1.0 and still an RFC.

    Any languageWorks with Mercurial, Jujutsu Open source, CC BY 4.0Source

  • Apache DevLake

    Pulls data from dev tools into one database and serves DORA dashboards.

    Self-hosted metrics where the team can add its own AI-authored field.

    Any languageWorks with GitHub, GitLab, Jenkins, Jira Open source, Apache-2.03.2k starsLatest release Sep 2026Source

  • Assisted-by trailer (Linux kernel policy)

    Kernel rule that AI-helped patches carry Assisted-by: AGENT:MODEL and a human sign-off.

    A precedent from a major project: label the tool and keep a human accountable.

    Any language Open sourcePart of Linux kernel

  • Co-Authored-By trailers

    A commit message trailer that names a co-author, which some agents use to mark AI work.

    The cheapest honest label, and GitHub already shows it on commits.

    Any languageWorks with GitHub, GitLab Open sourcePart of Git

  • DX

    Developer productivity platform that combines surveys and system data, with AI impact reports.

    Delivery numbers sit beside what developers say about working with the agents.

    Atlassian announced its acquisition of DX in September 2025.

    Any languageWorks with GitHub, GitLab, Jira Paid

  • Faros AI

    Engineering data platform with a community edition and AI adoption reports.

    Ties AI tool usage to delivery outcomes in one place.

    Any languageWorks with GitHub, GitLab, Jira, any CI Open source, paid tier, Apache-2.0Source

  • GitHub Copilot metrics and Claude Code analytics

    Vendor dashboards showing how much accepted code came from the assistant.

    First-party numbers for labeling agent share honestly, with the vendor's blind spots attached.

    The linked docs cover Copilot usage metrics only.

    Any languageWorks with GitHub Paid

  • Grafana Git AI dashboard

    Community Grafana dashboard that charts git-ai attribution data.

    The share of AI-written code lands on a graph next to delivery metrics.

    Its authors say to read it as adoption, not productivity.

    Any languageWorks with Grafana Open source

  • Jellyfish

    Engineering management platform that measures output, investment and AI tool impact.

    Compares work done with and without AI assistance across the whole org.

    Any languageWorks with GitHub, GitLab, Jira Paid

  • LinearB

    Delivery metrics plus automations that route and approve pull requests.

    Its automations can auto-approve small changes, which maps to rule 27.

    Any languageWorks with GitHub, GitLab, Bitbucket, Jira Paid

  • Middleware

    Open source DORA metrics dashboard deployed with Docker.

    A team gets DORA numbers on its own servers with no vendor contract.

    Any languageWorks with GitHub, GitLab, Jira Open source, paid tier, Apache-2.01.6k starsLatest release May 2025

  • Sleuth

    Tracks deploys and DORA metrics and links them to changes.

    Every deploy ties back to the change that caused it.

    Now split into Sleuth Skills, which governs agent skills, rules and MCP servers, and Sleuth DORA for metrics.

    Any languageWorks with GitHub, GitLab, Jira Paid

  • Swarmia

    Engineering metrics and working agreements for delivery teams, with AI adoption and cost tracking.

    Working agreements put a number on how fast reviews and merges happen.

    Any languageWorks with GitHub, GitLab, Jira, Linear Paid

Gone quiet 1

These tools are archived or shut down, or nobody has released or committed to them in two years. Some still work fine. Read the note before you adopt one.

  • DORA Four Keys

    Reference dashboard for the four DORA delivery metrics on BigQuery and Grafana.

    It shows the shape of the metrics and makes a frozen starting point.

    Archived in January 2024.

    Any languageWorks with GitHub, GitLab, Cloud Build, BigQuery Open source, Apache-2.02.2k starsLast change Jan 2024

Missing a tool that belongs here? Suggest it.

Every tool here was checked against its GitHub repository or its own site on 6 October 2026. Stars and dates come from GitHub on that day.