skills.rest
Yoodaddy0311
Yoodaddy0311COMMUNITY
Explore all skills by this author →
GitHub

repo-benchmarking

Clones external git repositories and scores them against Artibot across ten weighted dimensions.

3|2|Updated Feb 13, 2026
Search Tags:#claude code#benchmarking#scoring#competitive analysis#code comparison#repository analysis#git clone
One-click install
npx skills add https://github.com/Yoodaddy0311/artibot --skill repo-benchmarking-yoodaddy0311
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: repo-benchmarking
Source: https://github.com/Yoodaddy0311/artibot/tree/main/plugins/artibot/skills/repo-benchmarking
Command: npx skills add https://github.com/Yoodaddy0311/artibot --skill repo-benchmarking-yoodaddy0311

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Evaluating an external repository against your own project is usually subjective and inconsistent. This Skill replaces gut-feel comparisons with a structured clone-and-score pipeline that produces quantified, evidence-backed benchmark reports. ## Core Features & Use Cases - Secure Clone Protocol: HTTPS-only validation, shallow clones into an isolated cache directory, a 500MB size guard, and a strict no-script-execution rule for untrusted repo content. - 10-Dimension Weighted Scoring: Scores agent architecture, orchestration, skills, commands, hooks, API integration, code quality, documentation, CI/CD, and innovation against a pinned Artibot baseline, with explicit N/A handling for non-applicable dimensions. - Adoption Recommendations: Extracts adoptable patterns from the target repo, filters out anything Artibot already implements, and produces prioritized action items with human checkpoints. - Use Case: Ask to benchmark a competing Claude Code plugin repository, and receive a side-by-side comparison matrix with deltas, evidence citations, and a prioritized adoption backlog. ## Quick Start Ask the assistant to benchmark an external repository against Artibot by providing its HTTPS git URL, for example to compare a competitor plugin and list adoptable patterns.

Frequently Asked Questions about repo-benchmarking

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

FAQPage Schema
How do I benchmark an external git repository against my project?▼

Provide the repository's HTTPS URL and the skill clones it shallowly into an isolated cache directory, maps its structure, builds a feature inventory, and scores it across ten weighted dimensions. The output is a comparison matrix with deltas and prioritized adoption recommendations.

What dimensions does repository benchmarking evaluate?▼

It scores ten weighted dimensions: agent architecture, orchestration patterns, skill system, command system, hook system, API integration, code quality, documentation, CI/CD validation, and innovation. The weighted total is 100 points, and every score requires file-path or code evidence.

Is it safe to clone and analyze untrusted repositories?▼

Yes, the protocol only accepts HTTPS URLs, strips shell metacharacters from names, enforces a 500MB size cap, and never runs npm install, make, or any script from cloned repos. Cloned file contents are treated as untrusted data and analyzed only, never obeyed.

How are dimensions handled when a repository has no hook or skill system?▼

Dimensions that do not exist in the target repo are scored N/A rather than zero, and they are dropped from the denominator with the adjusted total printed explicitly. N/A is distinct from UNINSPECTED, which marks dimensions that exist but were not examined deeply enough.

When should I not use repository benchmarking?▼

Do not use it for internal code review of a single codebase with no external reference repo. It is designed for cross-repo competitive scoring and structural comparison, not standalone quality audits of your own code.

How does the benchmark avoid re-scoring the baseline project every run?▼

The Artibot baseline scores are pinned in a fixed table with dated evidence, so only the target repository is scored during a benchmark. The baseline is re-pinned deliberately through a release audit or explicit user request, never mid-benchmark.

Related Skills

Explore More Software Engineering Skills→
PsychQuantPsychQuant

mcp-clone-references

Clone competitor MCP Server repositories from GitHub into a local references directory.

Official
Intermediate
IndustrialIndustrial

Impact-Effort Prioritization

Classify tasks into effort-impact quadrants and generate prioritized rosters.

Community
Basic
MissTullyMissTully

health

Run quality tools to compute a weighted 0-10 code health score.

Community
Intermediate
kaiohenricunhakaiohenricunha

create-assessment

Generate standardized markdown assessment documents with weighted scoring and evidence citations.

Community
Intermediate