running-prompt-injection-eval

Community

Run prompt-injection evals end to end

Authorrocklambros
Version1.0.0
Installs0

System Documentation

What problem does it solve?

This Skill standardizes authorized prompt-injection evaluations for deployed LLM applications so teams can test defenses without inventing the workflow from scratch.

Core Features & Use Cases

It validates the pre-flight safety checks, ingests a user-supplied JSONL corpus, runs a structured dispatch and logging workflow, and classifies each response into blocked, passed, partial, or inconclusive. It is designed for pre-deployment safety reviews, guardrail audits, regression tests, and red-team engagements where corpus quality, scope control, and reproducible evidence matter. It also produces per-class summaries, false-negative and false-positive analysis, and a deduplicated finding pack ready for reporting.

Quick Start

Use the running-prompt-injection-eval skill to verify the signed RoE, validate the corpus and target endpoint, run the dry-run and full eval, and return the manifest, summary tables, and findings.

Dependency Matrix

Required Modules

None required

Components

references

💻 Claude Code Installation

Recommended: Let Claude install automatically. Simply copy and paste the text below to Claude Code.

Please help me install this Skill:
Name: running-prompt-injection-eval
Download link: https://github.com/rocklambros/rcs/archive/main.zip#running-prompt-injection-eval

Please download this .zip file, extract it, and install it in the .claude/skills/ directory.
View Source Repository

Agent Skills Search Helper

Install a tiny helper to your Agent, search and equip skill from 536,000+ vetted skills library on demand.