OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
Agent Evaluation & Observability

Harness Score

Harness Score is a TypeScript command-line developer tool that evaluates the repository-level guidance, skills, hooks, feedback systems, CI and hygiene surrounding AI coding agents. It reports a maturity level from L0 to L4, a score across six dimensions and concrete checks that identify gaps.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Harness Score — OpenViST project profile in Agent Evaluation & Observability.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use Harness Score when reviewing the controls around AI coding agents in a repository, planning improvements from concrete checks, or enforcing a selected maturity threshold in CI.

Who it is for

  • Teams maintaining repositories that use AI coding agents.
  • Developers who want a repeatable way to assess harness configuration across repositories.
  • Teams seeking a repository-level maturity gate in CI.

Documented uses

  • Measure the presence and structure of agent guidance, skills, hooks, feedback tools and safeguards.
  • Identify checks that block progress to the next harness maturity level.
  • Generate terminal, JSON, Markdown or badge-form results.
  • Use a minimum maturity level as a CI gate.
Before you choose

Requirements & limits

The project is documented as a CLI and is MIT-licensed. The README shows invocation through `npx harness-score` and references a GitHub Action.

  • The assessment focuses on repository files and configuration, not model performance or generated-code quality.
  • The README describes scoring as requiring no LLM calls or network access and producing consistent results for the same repository and commit.
  • The documented output includes a maturity level, dimensional breakdown and ranked checks.
  • The CLI includes JSON, Markdown and badge output options, and a minimum-level option for CI.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Score a repository and export a report

Assess a repository’s AI coding-agent harness and save the results in a documented format.

Before you start

  • A repository to assess.
  • The Harness Score CLI, invoked in the README with harness-score or npx harness-score.

Steps

  1. Run harness-score from the repository.

  2. For machine-readable output, run harness-score --json.

  3. To write a Markdown report, run harness-score --md report.md.

  4. To create a badge, run harness-score --badge harness-badge.svg.

Expected outcome

A maturity level and score with dimension-level results; the selected options provide JSON output, a Markdown report file or an SVG badge.

Guide limitations

  • The assessment checks repository artifacts and does not evaluate an agent’s generated code.
  • The documented commands describe scoring and report formats, not a complete remediation workflow.
Documentation 1 ↗
Source-based practical guide

Set a minimum maturity level for CI

Configure a CI workflow to fail when a repository does not meet the selected Harness Score level.

Before you start

  • A CI workflow for the repository.
  • The Harness Score GitHub Action referenced by the README.

Steps

  1. Add the paladini/harness-score@v1 action to the workflow.

  2. Set its min-level input to the required level, such as '3'.

Expected outcome

The workflow uses the configured minimum level as a gate; the README gives level 3 as an example.

Guide limitations

  • The example documents the action and its minimum-level input but does not provide a complete workflow configuration.
Documentation 1 ↗
COMPARE OPTIONS

Compare options for AI coding assistance

These projects document a common task and share a project role. This is a source-based comparison, not an execution benchmark or a guarantee of interchangeability.

ProjectDocumented purposePlatformAccess and requirements
Harness ScoreMeasure the presence and structure of agent guidance, skills, hooks, feedback tools and safeguards.CLIThe project is documented as a CLI and is MIT-licensed. The README shows invocation through `npx harness-score` and references a GitHub Action.
AgenttrailSee which project components and files are changing.CLI, DesktopKitchen setup requires Node.js 20 or later, a browser with WebGL and a local project folder. The README describes Kitchen as a public experimental preview.
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
14 October 2026
Original website title
paladini/harness-score · Official homepage title · retained in its source language
How we review sourcesSuggest a correction