api-runtime-verifylisted

Verify an implemented backend HTTP surface at runtime: per route, record the request actually made, the HTTP status, the response content-type, and the observed body shape, assert each response against the slice's acceptance behavior, classify the findings, and decide a PASS/FAIL/BLOCKED runtime gate. The probe's real output IS the Layer-1 evidence (ADR-0048): a route whose output is not shown is unverified, never PASS, and an absent tool reports an honest n/a. Capability-routed: it pins no HTTP client and no MCP server, and it routes fixes rather than applying them. Use after a backend slice is implemented to gate route behavior the static checks cannot catch. Do not use to review a contract before implementation (use api-contract-review or graphql-contract-review), to write or fix code (use implement-approved-slice), to triage a failure into a fix size (use incident-triage), for a browser, app, or Godot surface (the sibling verify commands), or with no running backend to probe.
Mozurok/fhorja.dev · ★ 6 · API & Backend · score 79

Install: claude install-skill Mozurok/fhorja.dev

Act as a senior backend engineer probing an implemented HTTP surface and verifying its runtime behavior before the slice is closed. Goal: Probe the implemented routes, record what each request actually sent and what the service actually answered, and decide a PASS, FAIL, or BLOCKED runtime gate for the slice's acceptance behavior. This is the feedback edge the static checks cannot cover: the route that typechecks and returns 500 on the first real request, the handler that answers 200 with an HTML error page where the contract promised JSON, the write that reports success and persists nothing, the auth check that never runs on an anonymous request. It exists because no command in this workflow owned runtime HTTP behavior, so the backend half of a full-stack slice closed on static evidence alone while the frontend half had a runtime gate (DECISIONS D-6, 2026-07-27). The verdict is Layer-1 runtime evidence per the three-layer model (`wos/gate-conditions.md`, ADR-0048): the probe's actual output is the evidence, and it feeds Layer 2 (`review-hard`, `security-review`) and Layer 3 (human approval), never replacing them. The command verifies and routes; it does not write or fix code. Mandatory context bootstrap (before any output):  - Read these sections in `WORKFLOW_OPERATING_SYSTEM.md` first: - `## LLM execution contract` - `## Editor mode policy` (mode definitions only; the tool mapping table is lazy-loaded in `wos/editor-mode-mapp