dottest-ai-agent-skills

This project contains skills and scripts that enable AI coding agents to automatically detect and fix static analysis violations reported by Parasoft dotTEST.

Overview

dottest-ai-agent-skills bridges the gap between dotTEST static analysis results and AI-powered remediation in your CI/CD workflow. Each skill packages prompt instructions and helper scripts so that a supported coding agent can run analysis, apply fixes, verify them, and optionally commit the changes — all without human intervention.

Skills interact with a Parasoft MCP (Model Context Protocol) server that exposes dotTEST analysis results to the agent in a structured, queryable format.

How It Works

1. Static analysis – Parasoft dotTEST analyses your .NET codebase and produces a violation report.
2. MCP server – The Parasoft MCP server reads the report and makes violations available to the AI agent via the Model Context Protocol.
3. AI agent skill – The skill instructions and scripts in this repository tell the agent how to query violations, apply code fixes, verify them with a build/test run and an incremental re-analysis, and commit the result.
4. CI/CD integration – The entire flow runs inside a CI/CD pipeline (GitHub Actions, Azure DevOps, Jenkins, etc.) with no interactive developer involvement.

Available Skills

Skill	Description
dottest-analyzer	Runs dotTEST static analysis on a .NET solution, collects violations, applies fixes, verifies each fix, and optionally commits changes. Supports scope limiting, Test Impact Analysis, branch-diff mode, and full CI/CD automation.

Prerequisites

Requirement	Notes
Parasoft dotTEST 2026.1+	Performs static analysis and unit-test verification on your .NET project.
Parasoft MCP server	Exposes dotTEST results to AI agents via MCP. Bundled with dotTEST under <DOTTEST_HOME>\integration\mcp\.
AI agent CLI	GitHub Copilot CLI or OpenAI Codex CLI (any MCP-compatible agent with SKILL.md support).
PowerShell	Execution policy set to AllSigned or RemoteSigned.

Getting Started

1. Register the MCP server with your agent (see the skill README for exact commands).
2. Copy the skill directory (e.g. dottest-analyzer/) into your agent's skills directory.
3. Configure required settings — at minimum DOTTEST_HOME and SOLUTION_PATH — via environment variables or a config file.
4. Run your pipeline – the agent will run analysis, apply fixes, and commit them as individual commits on the current branch.

See dottest-analyzer/README.md for full installation, configuration, and usage instructions.

Repository Structure

dottest-ai-agent-skills/
├── README.md                        # This file
└── dottest-analyzer/                # dottest-analyzer skill
    ├── SKILL.md                     # Skill instructions consumed by the AI agent
    ├── README.md                    # Full installation and usage guide
    ├── SKILL_DEMO.md                # Hands-on walkthrough / demo script
    ├── dottest-analyzer.config      # Annotated configuration file template
    ├── docs/
    │   └── scope_limitation.txt     # Pattern syntax for analysis scope limiting
    └── scripts/
        ├── dottest-analyze.ps1      # Runs dotTEST analysis
        ├── resolve-config.ps1       # Loads and validates configuration
        └── verify.ps1               # Verifies build and unit tests after a fix