harness-genome

Generates a 7-section repository readiness report via the metaharness genome command.

70.1k|8.4k|Updated Jun 2, 2025
One-click install
npx skills add https://github.com/ruvnet/claude-flow --skill harness-genome
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: harness-genome
Source: https://github.com/ruvnet/claude-flow/tree/main/plugins/ruflo-metaharness/skills/harness-genome
Command: npx skills add https://github.com/ruvnet/claude-flow --skill harness-genome

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Teams need a categorical, read-only assessment of a repository before scaffolding agent harnesses or publishing, but manually auditing repo type, risk, MCP surface, and test confidence is slow and inconsistent.

Core Features & Use Cases

  • 7-Section Genome Report: Returns repo_type, agent_topology, risk_score (0-1), mcp_surface, test_confidence (0-1), and publish_readiness (0-1) from npx metaharness genome <path> --json.
  • CI Risk Gating: The --alert-on-risk-above N flag exits with code 1 when the repo's risk score crosses a threshold, failing the build.
  • Graceful Degradation: Upstream exit codes 1 (needs-work) and 2 (blocked) are preserved as valid reports rather than treated as failures.
  • Use Case: Before minting a custom agent harness from a repo, run the genome report to confirm publish readiness is high and risk is low, then snapshot results over time to detect drift in agent topology recommendations.

Quick Start

Run the harness-genome skill against the current repository to produce a JSON readiness report with risk score and publish readiness.

Frequently Asked Questions about harness-genome

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I generate a repository readiness report with metaharness?

Run the harness-genome skill, which shells out to `npx metaharness genome <path> --json` with a 60-second timeout. It returns repo_type, agent_topology, risk_score, mcp_surface, test_confidence, and publish_readiness in JSON or markdown format.

How do I fail a CI build when repo risk is too high?

Pass the `--alert-on-risk-above N` flag, for example `--alert-on-risk-above 0.5`. The wrapper exits with code 1 when the repo's risk_score exceeds the threshold, which fails the CI build.

What is the difference between harness-genome and harness-score?

harness-score produces a 5-dimension numeric scorecard, while harness-genome produces a 7-section categorical and numeric report covering repo type, agent topology, risk, MCP surface, test confidence, and publish readiness.

Does the genome report modify the repository it scans?

No, the report is pure-read. It only inspects the target path and degrades gracefully, treating upstream exit codes 1 (needs-work) and 2 (blocked) as valid reports rather than errors.

Why does the genome command return a non-zero exit code?

Upstream metaharness exits 1 for a needs-work verdict and 2 for blocked. These are valid report outcomes, so the wrapper preserves them as `verdict` and `verdictExitCode` instead of failing.