{
  "context": "Tests whether the agent correctly chooses single-auditor mode for a narrow single-package target, produces a complete repo brief with all required fields, applies severity levels correctly, and places ambiguous items in Open Evidence Gaps rather than making unfounded claims.",
  "type": "weighted_checklist",
  "checklist": [
    {
      "name": "Single-auditor mode",
      "description": "Report explicitly states or clearly uses 'single-auditor' mode (not specialist-panel), given the narrow single-package scope",
      "max_score": 10
    },
    {
      "name": "Repo brief: product name",
      "description": "The Repo Brief section names the product or package (Ledger Service or similar)",
      "max_score": 5
    },
    {
      "name": "Repo brief: language/framework",
      "description": "The Repo Brief section identifies the implementation language (Go) and dependencies (database/sql, lib/pq, net/http)",
      "max_score": 5
    },
    {
      "name": "Repo brief: audit scope",
      "description": "The Repo Brief section explicitly states the audit scope (single package, named files or directory)",
      "max_score": 5
    },
    {
      "name": "Repo brief: docs checked",
      "description": "The Repo Brief section lists which docs were checked (README.md, go.mod, or similar)",
      "max_score": 5
    },
    {
      "name": "Repo brief: code surfaces checked",
      "description": "The Repo Brief section names the primary code surfaces inspected (e.g. handler.go, service.go, repository.go, model.go, middleware.go)",
      "max_score": 5
    },
    {
      "name": "Repo brief: module map",
      "description": "The Repo Brief section includes a top-level module map or file layout description",
      "max_score": 5
    },
    {
      "name": "Repo brief: test layout",
      "description": "The Repo Brief section describes the test layout and testing style (unit tests, mocked repo, skipped tests)",
      "max_score": 5
    },
    {
      "name": "Severity levels used",
      "description": "At least two findings include a 'Severity:' field with a value of critical, high, or medium",
      "max_score": 10
    },
    {
      "name": "Evidence from actual files",
      "description": "At least two finding blocks cite specific file names (e.g. service.go, repository.go) as evidence rather than speaking in generalities",
      "max_score": 10
    },
    {
      "name": "Open Evidence Gaps section present",
      "description": "Report contains a '## Open Evidence Gaps' section",
      "max_score": 10
    },
    {
      "name": "Ambiguous items in Open Evidence Gaps",
      "description": "At least one item is placed in '## Open Evidence Gaps' rather than stated as a confirmed finding (e.g. unknown production behavior, external system assumptions)",
      "max_score": 10
    },
    {
      "name": "Incremental recommendations",
      "description": "Recommendations describe incremental structural changes (NOT a full rewrite or microservice split)",
      "max_score": 15
    }
  ]
}