Has your agent write and run evals for Power Apps Code Apps and Generative Pages: feature checks, unit tests and a static security scan.
by Microsoft· AI & agentsUnclaimed · Yours? Claim it
4 images
About
Microsoft Power CAT's plugin that has your coding agent build and run an eval suite for a Power Apps Code App or a Generative Page. It reads your source code, maps what's built against what was planned, and shows the results in an HTML dashboard. The plugin is in preview. What its skills do - eval-generator-code-app: feature presence checks and a static security scan for a Power Apps Code App - eval-generator-gen-pages: feature presence checks, Vitest unit tests against a mocked dataApi, and a static security scan for a Generative Page in a model-driven app - Checks each planned feature against a requirements document (BRD) you give it, as a local file or a OneDrive or SharePoint link, or works the features out from the code - Flags risks such as hardcoded secrets, XSS, OData injection, PII in console logs and unsafe external calls, each with the file, line and how to fix it - Writes everything into an evals/ folder and adds npm run eval, which reruns every check and updates the dashboard - Keeps a snapshot of each run, so the dashboard shows which features improved or regressed How to install it In Claude Code, run /plugin marketplace add microsoft/power-cat-skills, then /plugin install powercat-procode-eval@power-cat-skills. Microsoft's guide also covers GitHub Copilot in VS Code. You need Node.js 18 or later with npm, and a Code App project (made with npx power-apps init) or a Generative Page. Good to know - Everything is static analysis on your own machine: it never connects to Dataverse or live connectors, and doesn't send your source code anywhere. - A describe-only mode shows what it would generate without writing any files. - It doesn't test runtime behaviour with real data, visual correctness, end-to-end flows, performance, deployment or accessibility, and it doesn't cover PCF components. - A clean security report means none of the known risk patterns were found, not that there are no vulnerabilities.
Once it’s added
“Generate code app evals for this project.”
“Eval my code app against this requirements document.”
“Check feature completeness of my code app.”
“Eval my generative page.”
“Test my generative page.”
“Review my genpage code.”
Before you add it
Reads only
Reads your Code App or Generative Page source code · Reads a requirements document (BRD) when you give it one
Makes changes · your agent can ask first
Writes an evals/ folder: checks, unit tests, runners and an HTML dashboard · Adds eval scripts to your package.json and runs npm install
From people who use it
Step by step
No reviews yet. Used Power CAT Pro-Code Eval? Say how it went.
Write a reviewRuns on your machine.
Runs on your machine.
/plugin marketplace add microsoft/power-cat-skills /plugin install powercat-procode-eval@power-cat-skills
Scanned and passed · checked 6 h ago
Nobody has shared how they use Power CAT Pro-Code Eval yet.
Plugins people install alongside Power CAT Pro-Code Eval.
Nothing new reaches your agent until a person has looked at it.
Scanned on every change, and again when our rules improve.
Claude CodeSupported
OpenClawSupportedTell us and a person on the trust team will look at it. Or email [email protected].
Plugins people install alongside Power CAT Pro-Code Eval.
Two debugging subagents and three commands that find the root cause of errors, set up error tracking and run multi-agent code reviews.
ByteDance's local MCP server that lets your agent run any shell command or script on your machine, with your own permissions.
Keeps your agent's plan, findings and progress in markdown files on disk, so long tasks survive /clear, crashes and compaction.
Security scan
Passed
Safe to install in every agent it supports. Claude plugins.
