Forward-looking movement triggers
Every score ships with a written-down answer: what would move this tool up, and what would move it down. Not 'watch for changes' — specific, falsifiable thresholds.
As of July 19, 2026
Six-dimension, evidence-graded scoring of agentic developer tools. We publish a written movement trigger for every score, so readers know what would move a tool up or down before it happens. Refreshed weekly, with caps and rationale public, and willing to flag negative signals that vendor-curated landscapes will not.
Most landscapes show you what exists. We show you what is shifting and what would shift it next. Three primitives keep the radar from drifting into commodity territory.
Every score ships with a written-down answer: what would move this tool up, and what would move it down. Not 'watch for changes' — specific, falsifiable thresholds.
Fourteen named, numbered caps mechanically bound dimension scores. When governance is capped at 12, the cap ID is listed and the rule is public. Auditable, not hand-wavy.
Every tool carries an Evidence Grade (A/B/C) reflecting hands-on vs. research depth. Vendor-only benchmarks get labeled. The grade tells you how much weight to put on the call.
Latest briefing
The single new thing this week isn't a model — it's a headcount. TCS committed to standing up **5,900–8,900 forward-deployed engineers** embedded with clients, explicitly naming OpenAI, Anthropic, and Microsoft as the competition. The "embed AI-native engineers with clients"...
Read full briefingParallel coordinates of all tools across the six dimensions
Browse and filter tools by signal level, category, or score
Industry timeline: model releases, funding, governance events
Programmatic access via the Model Context Protocol
Weekly briefings and the latest digest in long-form
Follow the radar
Weekly briefings and signal-level changes — check back Mondays.