Integrating Native Static Analyzers: detekt, SpotBugs, Rosly
Driving code analyzers from outside — source vs bytecode vs build-based, narrowing a general-purpose tool to one purpose, SARIF normalisation, and running them without touching the user's toolchain.
Free
Works with the AI Agents you already use
About this skill
Running a linter by hand is easy. Running five of them from a program, over code you did not write, on a machine you do not control, and getting back findings someone will act on — that is a different job.
This is what five integrations taught.
The split that decides everything is what the analyzer eats. Source (detekt, Bandit, ESLint) answers while the developer is typing. Compiled bytecode (SpotBugs) means the project must already be built and you have to find the output yourself. A full build (Roslyn) takes minutes and can fail for reasons unrelated to your scan — and still produce a valid report.
· Every general-purpose analyzer must be narrowed, and you only learn how by scanning somebody else's codebase. One default-on ESLint rule produced 31 of 33 findings on express; SpotBugs' non-security detectors produced 5 of 5 on spring-petclinic. · SARIF as the common format, and where each tool bends it — including the MSBuild escaped comma that silently downgrades Roslyn to SARIF v1 and makes every message parse as empty, with no error · Never build a .sln directly: every project writes to the same ErrorLog, only the last survives, and it is usually a test project · ESLint only lints under its working directory — point it elsewhere and you get zero findings and no error, the most confusing failure in this area · What an analyzer can see without type resolution, and why SpotBugs sees more but costs a build and dependency jars · Isolation: configuration on the command line, caches redirected, reports in temp, the user's project never modified
Per-analyzer reference with the exact invocation, narrowing flag, output format and traps for all five. Normalisation reference covering severity scales, rule ids, percent-encoded paths, and the CWE field that exists on some tools and hides inside the message on others.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
1 use
Times used to date
Free forever
No account required to browse
Trust & safety
Security scan
Security scan passed 29 days ago
- Free with an Agensi account