AL Griller Launches Static Analyzer for Business Central
Independent developer Yahyatouil releases AL Griller, a static analyzer for Dynamics 365 Business Central AL code built from 200+ files.

Stock photo for illustration only, not from the actual event
- Yahyatouil launches AL Griller, a static analyzer for AL code.
- Filtered rules from Microsoft's BCQuality repository with over 200 files.
- Shipped 26 rules in version one focusing on high reliability.
- Combines technical diagnostics with humorous roasts for developers.
Independent developer Yahyatouil has released AL Griller, a static analyzer designed for Dynamics 365 Business Central AL code that delivers humorous roasts whenever it detects real issues in the codebase. Behind the jokes lies a rigorous filtering process used to select the underlying rules from comprehensive knowledge repositories.
The project was built by working through Microsoft's BCQuality repository, a curated knowledge base and skills library dedicated to Business Central quality. Yahyatouil explicitly credits the BCQuality team while noting that AL Griller is an independent, community-run open-source project and not an official Microsoft product.

Stock photo for illustration only, not from the actual event
The BCQuality repository is extensive, containing over 200 knowledge files across 14 domains such as performance, query optimization, data modeling, error handling, events, and security. However, not every file could be converted into a static analysis rule, as many describe judgment calls that require broader context beyond a single-file text scanner.
The research pass acted as a strict filter applied file by file. Full reads were dedicated to performance with 54 files, along with query, data modeling, and error handling layers, while security utilized a representative sample of 8 out of roughly 20 files. Other domains like privacy and telemetry were flagged for future research.
"got summoned once per row like a tiny genie who can only grant one wish before vanishing back into the lamp, forever, on repeat"
Yahyatouil
Everything that passed the filter was classified as STRONG, GOOD, or EXPERIMENTAL, with shipped rules assigned P0 or P1 priorities. Exactly 26 rules made it into the final release from ALGRILL001 to ALGRILL026, while P2 and P3 rules were intentionally excluded to prioritize trust over rule count.
In software engineering, static analysis tools often suffer from high false-positive rates that lead to developer alert fatigue. AL Griller's deliberate choice to discard rules that lack deterministic detection mechanisms underscores a quality-first philosophy, ensuring that every warning generated is actionable and reliable.
Findings that required cross-object context beyond a single file's reach were explicitly rejected and documented in research files to assist future contributors. The AL Griller project remains active on GitHub, welcoming pull requests for new rules and roast improvements from the community.
Source: Dev.to
Found something wrong in this article? Report an issue with this article
Comments
Leave a Comment