TESIGN
SAVED

EN · KO

  1. TESIGN / RADAR
  2. 2026.09.17
  3. MartinLoop

MartinLoop

Execution-control CLI for coding agents

WHO USES ITA developer who hands tasks to a coding agent and has to review the results · A team that's been burned by an agent run costing more than expected · Anyone who wants to compare different agents or models under one governed flow

Wraps a coding agent with a spend cap and a verifier, so "I'm done" comes with an evidence-backed receipt instead of just a claim.

★ 130▲ +21 / 7dDaily star increases over 14 days. Hollow bars mean no data. The last day may be partial.2026-09-03 no data2026-09-04 no data2026-09-05 no data2026-09-06 no data2026-09-07 no data2026-09-08 no data2026-09-09 no data2026-09-10 no data2026-09-11 no data2026-09-12 no data2026-09-13 no data2026-09-14 no data2026-09-15 +212026-09-16 no datalargest one-day increase in the 14 days+21last 14 days · stars per dayhollow = no data

130 stars · checked on GitHub GitHub check 26 h late · 7-day +21 observed via GH Archive (as of )

MartinLoop's homepage headline: "Watch your AI burn $5.20. Then watch MartinLoop close the loop for $2.30."
MartinLoop's homepage headline: "Watch your AI burn $5.20. Then watch MartinLoop close the loop for $2.30."
MartinLoop's homepage headline: "Watch your AI burn $5.20. Then watch MartinLoop close the loop for $2.30."

TESIGN TAKE

It doesn't try to make the agent smarter — it treats uncontrolled execution by an already-capable agent as the actual problem, which is a sharper framing than most.

TRY

AT A GLANCE

LICENSE
Apache-2.0
USAGE
Commercial use, redistribution OK. Keep notices; mark changes.
LANGUAGE
TypeScript
PLATFORM
cli
ACTIVITY
last commit 1 day ago () · latest release [unconfirmed]
COMMUNITY
contributors [unconfirmed] · open issues [unconfirmed] · made by: [unconfirmed] Reference time unavailable
SOURCES
GitHub
OPEN SOURCE
YES
FIRST SEEN
CATEGORY
DEV TOOLS · AI

A summary, not legal advice.

WHY IT MATTERS

A task that looks like a small fix can turn into dozens of retries, a blown budget, and a diff nobody trusts. MartinLoop wraps a coding agent (Claude Code, Codex, and others) with one execution contract — objective, verifier command, budget, iteration cap — around the run. Completion only fires when fresh verifier evidence is bound to that specific run and workspace ("VERIFIED" means the configured checks passed, not that the code is bug-free), and every run leaves a receipt you can re-inspect with martin dossier.

BUILD FROM THIS

  • Put a hard budget and attempt cap on a task handed to a coding agent
  • Bind an independent verifier command (e.g. npm test) so completion needs passing evidence, not just a claim
  • Compare outcomes across different agents using the same receipt format

WHO IT'S FOR

A developer who hands tasks to a coding agent and has to review the results
A team that's been burned by an agent run costing more than expected
Anyone who wants to compare different agents or models under one governed flow

START IN 5 MINUTES

# Try it now:
$ npx -y martin-loop@latest start
$ npx -y martin-loop@latest demo
$ cd martin-loop-demo && npm install
$ npx -y martin-loop@latest run "describe the task" --verify "npm test" --budget-usd 2 --max-iterations 1
$ npx -y martin-loop@latest dossier --latest

CAVEATS

  • Apache-2.0 · needs Node.js 20+
  • The hosted dashboard and team features are still early access (per the official homepage)
  • The README states plainly that VERIFIED means the configured checks passed, not that the code is safe to merge

RECEIPT

FIRST SEEN
AT SOURCE
KEPT
CREATED → FIRST SEEN
164d
WHEN FIRST SEEN
130 ★
SOURCES
GitHub

The same facts in machine-readable form — View as Markdown · JSON

SIMILAR TOOLS

Up to five from DEV TOOLS by ★ total: edited entries first, then repository cards; this entry is highlighted. Drawn from the same stored snapshot as the rankings — a comparison, not a recommendation.

IMAGENAME★ TOTAL▲ 7dLICENSEPLATFORMLAST PUSH
ECC README banner: headline 'The operating system for AI agent harnesses', harness chips, and skill, agent and command lists.
eccSkills, memory and security for agents★ 259k▲ +832MITpermissivewindows · macos · linux · cli
DeepSeek Harness landing: blue background, the headline 'Everything is a plugin' and a quick-start box with the npx command.
DeepSeek HarnessPlugin agent harness for developers★ 225k▲ +1,356MITpermissiveweb · cli
ponytail.dev landing page: a line-drawn developer with a ponytail on black, the title 'ponytail' and an install button.
ponytailMake your agent write less code★ 139k▲ +1,068MITpermissive—
Spec Kit docs landing page — the headline 'Build with a spec, fix a bug, or assess an idea' and the uv tool install specify-cli command
Spec KitSpec-first process for coding agents★ 137k▲ +560MITpermissivecli · windows · macos · linux
MartinLoop's homepage headline: "Watch your AI burn $5.20. Then watch MartinLoop close the loop for $2.30."
MartinLoopExecution-control CLI for coding agentsTHIS ENTRY★ 130▲ +21Apache-2.0permissivecli

ALSO WORTH SEEING

Spec KitSpec-first process for coding agentsSpec Kit docs landing page — the headline 'Build with a spec, fix a bug, or assess an idea' and the uv tool install specify-cli commandInstall to use
★ 137k▲ +560 / 7d
Makes your coding agent write the spec first, then plan, build and verify against it.HindsightNarrative map for YouTube viewersHindsight landing page: thousands of video dots grouped into issue clusters, with a central search box for a topic, creator or YouTube link.Use in the browser
★ 0▲ 1 HN
Remembers 49,411 videos, maps 22,904 by issue, traces how a YouTuber's argument changed.Qwen2.5 Program Slicing Security AuditorJavaScript security audit for web devsThe auditor's dark UI: a four-step pipeline diagram at the top, a code editor with the 'Run 4-Step Security Audit' button on the left, and the findings panel on the right.Use in the browser
♥ 1
Paste JS or HTML; a fine-tuned 7B model traces each risky sink to an exploit and verdict.FormaONNX graph editor for ML engineersForma editor: a sample ONNX graph on the canvas with the Conv node selected and its editable attributes, tensor shapes and parameter count in the Layer Inspector.Use in the browser
★ 3▲ 2 HN
Drop an .onnx file in the browser, see the graph, edit a node and export — no Python.

TIMELINE

  1. Repository created
  2. FIRST SEEN BY TESIGN
  3. Published on TESIGN

SAME DAY

  1. 01

    BG0

    No-account background remover

    Removes a photo's background entirely inside your browser — no upload, no server, no account.

    Use in the browser
    ★ 79▲ +28 / 7d
  2. 02

    ZapFast

    Lightweight native WhatsApp client

    A from-scratch Rust WhatsApp client with no browser engine — about 150MB idle RAM on Linux, versus 1.13GB for WhatsApp Web.

    Install to use
    ★ 107▲ +39 / 7d
  3. 03

    Wealthfolio

    Local-first personal finance tracker

    An open-source portfolio tracker that keeps investments, net worth and spending entirely on your device — no cloud account.

    Install to use
    ★ 9k▲ +15 / 7d
  4. 04

    page-mascot

    Cursor-following page mascot

    A tiny React component that puts a character on your page which turns its head to follow the cursor and reacts when clicked.

    For developers
    ★ 307▲ +39 / 7d
  5. 05

    Podroid

    Linux container VM for your phone

    Boots a real, separately-kernelled Alpine Linux VM on an unrooted Android phone, running Podman, Docker and LXC exactly as they behave on a server.

    Install to use
    ★ 2.6k▲ +14 / 7d
  6. 06

    Capsule

    Single-file app player

    A desktop player that packs an HTML app and its data into one SQLite file — no account, no server, just a file to share.

    Install to use
    ▲ 333 HN
  7. 07

    SparkyFitness

    Self-hosted family health tracking

    A self-hosted health tracker for nutrition, exercise, sleep, fasting, mood and body metrics — all on a server you run yourself.

    Self-host
    ★ 6.1k▲ +13 / 7d
  8. 08

    Inkstone

    Self-hosted notebook on Cloudflare

    A self-hosted Markdown notebook you deploy on your own Cloudflare Workers account, with realtime sync, search and backlinks built in.

    Self-host
    ★ 819▲ +10 / 7d
  9. 10

    OpenHunterAI

    Local AI red-team workspace

    An alpha-stage local workspace that runs attacker-style security assessments — strictly against public targets you own or are authorized to test.

    Self-host
    ★ 165▲ +15 / 7d