uinaf/docs
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 boot/readiness infrastructure setup.
95
95%
Does it follow best practices?
Impact
98%
1.06xAverage score across 4 eval scenarios
Passed
No known issues
Evaluation results
100%
Fix Documentation Drift After a Backend Refactor
bootstrap.sh replaced
100%
100%
config path updated
100%
100%
api-v2.md link fixed
100%
100%
API version updated in runbook
100%
100%
PDF export section removed
100%
100%
test-pdf.sh reference removed
100%
100%
Audit report produced
100%
100%
Audit covers all stale refs
100%
100%
Audit includes verification
100%
100%
Valid links preserved
100%
100%
100%
26%Keep Local Discovery Out of Checked-In Docs
Repo-owned setup used
100%
100%
Repo-owned verification used
100%
100%
Canonical setup linked
100%
100%
No absolute paths or local helper
100%
100%
No host or user identity
62%
100%
No private sibling or dashboard
16%
100%
No account profile
20%
100%
Source boundary reported
50%
100%
Promotion requires approval
100%
100%
No invented policy
100%
100%
92%
-2%Onboard a New AI Agent to the Payments Service
Length constraint
100%
100%
Boot command present
100%
100%
Test command present
100%
100%
Links to deep docs
80%
100%
No directory listing
100%
20%
No auto-generated content
100%
100%
No full API docs
100%
100%
Deprecated script removed
60%
100%
Key deviating conventions
100%
100%
Doc report produced
100%
100%
100%
Restructure Documentation for an Open Source CLI Tool
README hero section
100%
100%
README install near top
100%
100%
README links out to CONTRIBUTING
100%
100%
README links out to SECURITY
100%
100%
No content duplication
100%
100%
CONTRIBUTING setup first
100%
100%
CONTRIBUTING validation section
100%
100%
SECURITY private-first
100%
100%
SECURITY contact present
100%
100%
SECURITY short
100%
100%
Doc changes summary
100%
100%
- Evaluated
- Agent
- Claude Code
- Model
- Claude Sonnet 4.6