What problem does it solve? Technology decisions often stall on unverified claims or end in hedged "it depends" answers with no citation trail. This Skill forces a structured investigation of a candidate library, dependency, or file format against a fixed question set, validates the result against a JSON schema, and produces exactly one actionable recommendation. ## Core Features & Use Cases - Structured research workflow: Dispatches a research subagent with a fixed question set (correctness, integration_fit, footprint, maturity, practical_fit) and requires cited evidence for every verdict. - Schema validation gate: Validates the subagent's YAML output with scripts/validate-tech-evaluation.sh before any finding is rendered, rejecting empty evidence, out-of-enum verdicts, or thin rationales. - Finding record generation: Renders validated results into a dated FINDING record via the context-file skill and regenerates the context index for discoverability. - Use Case: Choosing between mustache, micromustache, and a hand-rolled templating function for a Lambda-bundled project — the Skill verifies the actual Terraform bundling config rather than trusting assumptions, then delivers a single adopt/reject verdict with citations. ## Quick Start Ask the agent to properly evaluate a specific npm package or file format for your integration point, naming the alternatives already considered.
