Stress and pressure

Introducing the evlog CLI. Lighthouse for observability. It reads your source, understands your framework, finds every entry point, and maps what context you'll have when things break. evlog map Scans your app and gives you an observability score. Finds handlers logging nothing, routes with no audit trail. Tells you the exact three to fix first. evlog init Guided setup that reads your project before asking anything. Detects your framework, finds repeated errors across files, generates error catalogs seeded with your actual code. Patches configs at AST level so nothing breaks. evlog doctor Verifies evlog is installed and writing where you expect. Every command outputs JSON. Every check can be disabled in code. Gates CI with --min-score and --baseline to catch regressions per requirement. A CLI that understands your codebase, knows which endpoints exist, what they do, where the gaps are. That's exactly what AI agents need to reason about observability. The static analysis, framework detection, deterministic scans, structured output. This is the foundation agents need to understand your app's coverage and make informed decisions. Fair warning: this is complex tooling. I've tested it thoroughly, everything works in theory, but it's early. Bugs might surface. Shipping anyway because like we say at Vercel, ITG! evlog.dev/cli