skills.rest
EndUser123
EndUser123COMMUNITY
Explore all skills by this author →
GitHub

ai-pi-nv-deepseek

Analyze a target file with Nvidia's DeepSeek via Pi to produce a JSON report.

Updated Apr 26, 2026
Search Tags:#deepseek#code-review#pi#vulnerability-analysis#adversarial-review#nvidia-nim
One-click install
npx skills add https://github.com/EndUser123/cc-marketplace --skill ai-pi-nv-deepseek
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: ai-pi-nv-deepseek
Source: https://github.com/EndUser123/cc-marketplace/tree/main/plugins/cc-skills-ai-cli/skills/ai-pi-nv-deepseek
Command: npx skills add https://github.com/EndUser123/cc-marketplace --skill ai-pi-nv-deepseek

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This skill enables automated adversarial review of code and text by invoking Nvidia's DeepSeek via Pi to produce a structured feedback report that highlights vulnerabilities and issues.

Core Features & Use Cases

  • Parse a target file and run DeepSeek-V3.2 via nvidia-nim to evaluate it and generate a deterministic JSON payload.
  • Return a JSON object with a score (0-1), a one-sentence summary, and a list of issues detected.
  • Provide actionable remediation guidance and a concise findings report suitable for engineering reviews.

Quick Start

Run pi with the model nvidia-nim/deepseek-ai/deepseek-v3.2 on the target file to obtain a JSON score, summary, and issues.

Frequently Asked Questions about ai-pi-nv-deepseek

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

FAQPage Schema
How do I automate adversarial code review using DeepSeek?▼

This Skill runs DeepSeek V3.2 via Nvidia NIM to perform adversarial code review on a target file, returning a deterministic JSON report. The output includes a 0-1 score, a one-sentence summary, and a list of detected issues.

What does the structured JSON output contain when analyzing a file for vulnerabilities?▼

The structured JSON output from a vulnerability analysis contains a score from 0 to 1, a one-sentence summary, and a list of identified issues. This provides concise findings suitable for engineering reviews and actionable remediation guidance.

Can I use the pi tool with nvidia-nim to evaluate multiple files and modules?▼

Yes, you can use the pi tool with nvidia-nim to evaluate multiple files and modules. The process is applicable to codebases or documents requiring rapid adversarial assessment, generating deterministic output per run for each target file.

Do I need specific inputs to run vulnerability analysis with DeepSeek V3.2?▼

Yes, you need a valid target file to run vulnerability analysis with DeepSeek V3.2. The system requires invoking the pi tool with the deepseek-v3.2 model and explicitly handles potential missing inputs or rate limits during the parsing process.

Why does my adversarial review output change between runs on the same codebase?▼

Your adversarial review output should not change between runs because the system generates deterministic output per run. If variations occur, verify that you are correctly parsing the JSON output from the pi tool using the specified deepseek-v3.2 model.

Related Skills

Explore More Software Engineering Skills→
EndUser123EndUser123

ai-pi-nv-nemotron-ultra

Run adversarial code and prompt reviews via Nvidia Nim with Llama 3.1 Nemotron Ultra.

Community
Advanced
Logisoft-webLogisoft-web

vulnerability-scanner

Scan code repositories for vulnerabilities in dependencies, secrets, and configurations.

Community
Advanced
hardw00thardw00t
90

sast-orchestration

Orchestrate Semgrep, CodeQL, and Bandit scans across codebases.

Community
Advanced
CyapstayeCyapstaye
9

evaluating-llms-harness

Benchmark large language models across academic and reasoning tasks.

Community
Advanced