harness-status

Inspects Harness project assets, hook wiring, baseline state, and task progress to report a health snapshot.

Updated May 15, 2026
One-click install
npx skills add https://github.com/Alan-IFT/harness-kit --skill harness-status-alan-ift
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: harness-status
Source: https://github.com/Alan-IFT/harness-kit/tree/main/skills/harness-status
Command: npx skills add https://github.com/Alan-IFT/harness-kit --skill harness-status-alan-ift

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Teams using the Harness Engineering toolkit lose track of whether their project is correctly set up: required documents may be missing, safety hooks may be dangling or miswired, and verification results go stale. This Skill produces a read-only, one-page status report so you can see the project's Harness health at a glance without running any tests. ## Core Features & Use Cases - Asset inventory: Checks 14 required assets (CLAUDE.md, docs/workflow.md, docs/tasks.md, baseline.json, guard-rm.sh, and more) and reports each as present or missing. - Hook wiring diagnosis: Resolves the effective hook source across .claude/settings.local.json and .claude/settings.json, then classifies the guard-rm PreToolUse hook as installed, dangling, malformed, absent, or opted out, plus per-event congruence for Stop, UserPromptSubmit, and SessionStart hooks. - Health score and recommendations: Computes a 12-point health score from assets, baseline freshness, last verify result, and task state, then lists concrete fix steps such as running /harness-init or /harness-upgrade. - Use Case: After cloning a Harness-enabled repository onto a new machine, run the status check to discover that lifecycle hooks were never installed locally and get the exact remediation command. ## Quick Start Ask the AI to show the current Harness status of this project, including which assets are missing and whether the safety hook is wired.

Frequently Asked Questions about harness-status

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

FAQPage Schema
How do I check the Harness health of my Claude Code project?▼

Run the harness-status skill to get a one-page snapshot covering 14 required assets, baseline state, last verify result, active tasks, and a 12-point health score. It is read-only and never runs tests or modifies files.

How do I verify the guard-rm safety hook is wired correctly?▼

The status report resolves the effective hook source from .claude/settings.local.json or .claude/settings.json, then checks whether a PreToolUse entry references guard-rm.sh and whether every extracted script path exists on disk. Verdicts range from installed and wired to dangling, malformed, absent, or opted out.

Why does harness-status report my hooks as not installed on this machine?▼

Hooks live in settings files that may be machine-local. If neither .claude/settings.local.json nor .claude/settings.json declares lifecycle hooks, the report shows never-installed for this clone. Run the install-hooks script to wire them on the current machine.

Does harness-status run the test suite or modify my project?▼

No. The skill is strictly read-only: it reads settings files, docs/tasks.md, baseline.json, and verification_history.log but never executes verify_all or writes anything. Use /harness-verify if you want a fresh verification run.

What does a dangling hook wiring mean in the status report?▼

Dangling means a hook command references a script path that does not exist on disk, or contains an unresolved placeholder token. The report prints the exact command and missing path, plus a fix line such as running /harness-upgrade for committed wiring.