capture-evidence

Generate harness, metric, splits, and baseline artifacts from workloads.

10|5|Updated Jun 3, 2026

Search Tags:#evaluation #baseline #artifact #workload #harness #local-first #capture-evidence

Ready to run

npx skills add https://github.com/understudylabs/understudy-agent-tools --skill capture-evidence

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: capture-evidence
Source: https://github.com/understudylabs/understudy-agent-tools/tree/main/skills/capture-evidence
Command: npx skills add https://github.com/understudylabs/understudy-agent-tools --skill capture-evidence

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Turn real-world workloads into auditable local evidence by generating repeatable harness, metric, splits, and baseline artifacts to establish a baseline before changes.

Core Features & Use Cases

Auditable artifact generation: Produce harness.json, environment.json, metric.json, splits.json, and baseline.json from a workload.
Local-first evaluation workflow: Enables offline, private benchmarking without hosted services or accounts.
Baseline-driven optimization: Provides a traceable path to compare current vs. future model performance.

Quick Start

Run the understudy capture-evidence workflow against your repository to generate local baseline artifacts and a repeatable evaluation.

Frequently Asked Questions about capture-evidence

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

FAQPage Schema

How do I establish an offline evaluation baseline for real-world workloads?▼

To establish an offline evaluation baseline, generate repeatable harness, metric, splits, and baseline artifacts from your workload. This creates a traceable local foundation to compare current versus future model performance before making changes.

What is local-first benchmarking and do I need existing traces to use it?▼

Local-first benchmarking enables private, offline evaluation without hosted services or accounts. You do not need existing traces or baselines; the workflow generates initial baseline artifacts directly from your repository.

How do I generate auditable artifacts with provenance for model optimization?▼

Generate auditable artifacts by running a capture-evidence workflow against your repository. It produces timestamped files like harness.json, environment.json, and baseline.json with source refs to enforce provenance for baseline-driven optimization.

Can I run private workload profiling without hosted services or accounts?▼

Yes, you can run private workload profiling without hosted services. The local-first workflow operates entirely offline, generating environment.json and workload-profile.md files to support private benchmarking.

What's the best way to compare current vs future model performance locally?▼

The best way to compare model performance locally is by establishing a baseline using generated splits.json and baseline.json artifacts. These provide a traceable path to evaluate future changes against your original workload.