Update repo documentation and agent-facing guidance such as AGENTS.md, README.md, docs/, specs, plans, and runbooks. Use when code, skill, or infrastructure changes risk doc drift or when documentation needs cleanup or restructuring. Do not use for code review, runtime verification, or `agent-readiness` setup.
98
100%
Does it follow best practices?
Impact
94%
1.00xAverage score across 3 eval scenarios
Passed
No known issues
{
"context": "Tests whether the agent systematically detects and fixes all stale references in documentation after a code refactor, verifying actual file/path existence before making changes, and produces a structured audit report documenting every issue found.",
"type": "weighted_checklist",
"checklist": [
{
"name": "bootstrap.sh replaced",
"description": "Both AGENTS.md and README.md no longer reference `scripts/bootstrap.sh` — replaced with `scripts/setup.sh` or removed",
"max_score": 12
},
{
"name": "config path updated",
"description": "README.md no longer references `config/app.yaml` or `config/app.yaml.example` — updated to `config/settings.yaml` variants",
"max_score": 12
},
{
"name": "api-v2.md link fixed",
"description": "AGENTS.md and README.md no longer link to `docs/api-v2.md` — updated to `docs/api-v3.md`",
"max_score": 10
},
{
"name": "API version updated in runbook",
"description": "docs/ops/runbook.md no longer references `/api/v2/health` or `/api/v2/webhooks/stripe` — updated to v3 equivalents",
"max_score": 10
},
{
"name": "PDF export section removed",
"description": "docs/ops/runbook.md no longer contains the PDF export troubleshooting section (feature was removed)",
"max_score": 10
},
{
"name": "test-pdf.sh reference removed",
"description": "No documentation file references `scripts/test-pdf.sh` after the fix",
"max_score": 8
},
{
"name": "Audit report produced",
"description": "doc-audit.md file exists with a structured list of stale references found",
"max_score": 8
},
{
"name": "Audit covers all stale refs",
"description": "doc-audit.md mentions at least 4 distinct stale references that were found and fixed",
"max_score": 10
},
{
"name": "Audit includes verification",
"description": "doc-audit.md states what the correct replacement is and confirms it was verified to exist (not just assumed)",
"max_score": 10
},
{
"name": "Valid links preserved",
"description": "Links that are still valid (docs/architecture.md, docs/ops/deploy.md, docs/ops/runbook.md, src/utils/logger.ts, src/lib/calculator.ts) are NOT removed from docs",
"max_score": 10
}
]
}