tech-eval-researchlisted

# Technology Evaluation Research Produce a decision-grade comparative evaluation of technology candidates: hard filters that eliminate, capability checklists resolved per-feature from primary sources, an operationalized maturity assessment, adversarial verification of the claims that decide the ranking, and a standard report with a comparison matrix and a recommendation. The core insight this process is built on: **vendors put the paywall where the docs are vaguest**. Marketing pages name features; pricing pages gate them; only the LICENSE file, the docs, and the code tell the truth. So every verdict comes from primary sources, every claim that decides a ranking gets adversarially re-checked, and "exists" is never confused with "exists in the tier the user can actually use." ## Process overview ``` Phase 0 Scope the brief (criteria, candidates, deliverable — with the user) Phase 1 Parallel research (subagents, primary sources only) Phase 2 Adversarial verify (try to refute the decision-critical claims) Phase 3 Synthesize & rank (filters eliminate, biases down-rank) Phase 4 Deliver (report always; deck on request) ``` Track phases in the task list. Phases 1 and 2 are subagent work; keep synthesis and ranking judgment in the main session. If subagents aren't available in the current environment, run the same phases sequentially inline — the source discipline and verification pass matter more than the parallelism. Scale effort to the