I was reading through ai-code-trust-validator and noticed an interesting split in the roadmap.
The README frames it as a broad “quality gate” for AI-generated code: security scanning, hallucination detection, logic validation, SARIF/HTML/PDF reports, IDE integrations, API server, watch mode, etc. But the open issues include both big surface-area expansions like PHP / Rust / Java/Kotlin support and more architectural pieces like configurable severity thresholds, custom rule profiles, --baseline, --git-diff, and incremental validation.
My outside read is that custom rule profiles / severity thresholds might be the cleaner contributor seam than “add another language” at this stage. It forces the rule model to become explicit without requiring a new parser/analyzer stack, and it probably makes later language support less ad hoc.
Genuine question, not a pitch: when you think about getting outside contributors involved here, do you find yourself trying to point people at visible language-support issues, or at core seams like rule profiles / baseline / diff mode first?
A short reply is plenty.
— Ray, a founder working in an adjacent devtools/recruiting space.
I was reading through
ai-code-trust-validatorand noticed an interesting split in the roadmap.The README frames it as a broad “quality gate” for AI-generated code: security scanning, hallucination detection, logic validation, SARIF/HTML/PDF reports, IDE integrations, API server, watch mode, etc. But the open issues include both big surface-area expansions like PHP / Rust / Java/Kotlin support and more architectural pieces like configurable severity thresholds, custom rule profiles,
--baseline,--git-diff, and incremental validation.My outside read is that custom rule profiles / severity thresholds might be the cleaner contributor seam than “add another language” at this stage. It forces the rule model to become explicit without requiring a new parser/analyzer stack, and it probably makes later language support less ad hoc.
Genuine question, not a pitch: when you think about getting outside contributors involved here, do you find yourself trying to point people at visible language-support issues, or at core seams like rule profiles / baseline / diff mode first?
A short reply is plenty.
— Ray, a founder working in an adjacent devtools/recruiting space.