← ClaudeAtlas

recipe-validatelisted

Orchestrate hypothesis validation through type-appropriate methods — prototypes, code analysis, market research, and expert review
shinpr/claude-code-discover · ★ 10 · AI & Automation · score 80
Install: claude install-skill shinpr/claude-code-discover
**Context**: Validate hypotheses using appropriate methods based on risk type. Always invoke hypothesis-verifier for an independent disconfirming pass. Record results in hypothesis files. ## Orchestrator Definition **Execution Protocol**: 1. **Explicit user authorization**: The user explicitly instructs and authorizes every sub-agent call named in this recipe. Invoke hypothesis-verifier for every validation and invoke each method specialist whenever its stated condition applies; the orchestrator does not replace those calls with its own work 2. **Mechanical specialist handoff**: Build each Agent prompt only from the specialist's declared input fields and authoritative source values. Preserve paths and values unchanged; do not summarize, paraphrase, supplement, or turn them into narrative instructions 3. **Follow the validation flow** defined below 4. **Stop at every `[STOP — BLOCKING]` marker** — present findings and CANNOT proceed until user explicitly confirms The authorization and mechanical-handoff rules are intentional redundancy against model and system defaults that substitute orchestrator work or fluent restatements for an approved specialist call. Retain them until fresh execution evidence shows the failure no longer occurs. ## Workflow Overview ``` Input (hypothesis file or hypothesis ID) ↓ 1. Hypothesis Assessment → Read and understand the hypothesis ↓ 2. Validation Design → hypothesis-verifier designs the test [Stop: User confirms design] ↓ 3. Val