sddl-eval-init

Initializes a local sdd-lite-evals workspace for a project and active CLI provider.

Updated Apr 2, 2026
One-click install
npx skills add https://github.com/nico0695/ai-tools --skill sddl-eval-init-nico0695
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: sddl-eval-init
Source: https://github.com/nico0695/ai-tools/tree/main/sdd/sdd-lite-evals/skills/sddl-eval-init
Command: npx skills add https://github.com/nico0695/ai-tools --skill sddl-eval-init-nico0695

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Setting up an sdd-lite evaluation workspace requires resolving the harness location, detecting the project root and active CLI provider, and installing evaluator skills correctly. This Skill automates that initialization while validating the existing sdd-lite setup without modifying it. ## Core Features & Use Cases - Harness Resolution: Locates the real skill directory and verifies the presence of scripts/sddl_eval.py, falling back to .harness-root for copied installations. - Environment Inspection: Detects the project root, active provider (codex, claude, or opencode), Git state, installed wrappers, and chat language before asking minimal questions. - Controlled Initialization: Runs sddl_eval.py init with workspace id, install method (symlink or copy), and language, then reports workspace path, installed evaluator skill paths, sdd-lite hash, and Level 1 findings. - Use Case: Before running your first sdd-lite evaluation in a repository, use this Skill to create the project-local evaluation control data and install evaluator skills via symlink without touching the sdd-lite installation itself. ## Quick Start Initialize the sdd-lite-evals workspace for this project using the current CLI provider before running any evaluations.

Frequently Asked Questions about sddl-eval-init

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

FAQPage Schema
How do I initialize an sdd-lite-evals workspace for my project?▼

Run the initialization flow, which resolves the harness directory, inspects your project root and active provider, then executes sddl_eval.py init with your workspace id, install method, and language. It reports the workspace path and installed evaluator skill paths on success.

Which CLI providers does sdd-lite evaluation support?▼

The initialization supports codex, claude, and opencode as provider options. The active provider is inferred during inspection, and you only need to confirm it or specify a non-default sdd root if the inference is unresolved.

Should I use symlink or copy installation for evaluator skills?▼

Symlink installation is recommended because it keeps evaluator skills linked to the harness source. Copy installation is available as an alternative, in which case the harness is resolved through the .harness-root file.

Why does sdd-lite-evals initialization return blocked?▼

A blocked result means the sdd-lite setup is missing or incompatible, or the harness could not be resolved to an existing installation. The Skill explains the issue and stops without repairing the project, injecting wrappers, or calling sddl-init.

Does sddl-eval-init install or repair sdd-lite itself?▼

No. It validates the existing sdd-lite installation but never installs, repairs, or modifies it. If the setup is missing or incompatible, initialization stops with a blocked status and an explanation.