ashfordeOU

Setuhq Benchmarks — Data skill for Claude Code

Data community

Report how well SetuHQ's reviewer panel and pipeline perform, from the stored benchmarks (real TOIs, injection-recovery, pi Men validation, calibration), via the SetuHQ MCP server.

How to install Setuhq Benchmarks

Installs to ~/.claude/skills/ashfordeou-setuhq-setuhq-benchmarks/SKILL.md

Terminal

mkdir -p ~/.claude/skills/ashfordeou-setuhq-setuhq-benchmarks && curl -fsSL https://raw.githubusercontent.com/ashfordeOU/SetuHQ/HEAD/commands/setuhq-benchmarks.md -o ~/.claude/skills/ashfordeou-setuhq-setuhq-benchmarks/SKILL.md

Restart Claude Code, or start a new session, for it to be picked up.

What Setuhq Benchmarks does

description: Report how well SetuHQ's reviewer panel and pipeline perform, from the stored benchmarks (real TOIs, injection-recovery, pi Men validation, calibration), via the SetuHQ MCP server argument-hint: "[what to check, e.g. 'completeness on real TOIs', 'how independent are the Sentinels', 'which providers can run here']"

What the benchmarks say

Request: **$ARGUMENTS**

Do this:

Call benchmark_results with the one that answers the question: real_tois (the blind bench

Alternatives in Data

LLM Redteam — LLM red-team corpus runner — fires categorized prompt-injection / jailbreak / system-prompt-leak / data-exfil 4.5k ★
Claude Deep Research Skill — Enterprise-grade deep research skill for Claude Code with 8-phase pipeline, source credibility scoring, and au 1k ★
Vibesec — Helps write secure code by preventing common vulnerabilities including IDOR, XSS, SQL injection, SSRF, and wea 687 ★