Judge covered / partial / missing per requirement with verbatim evidence.
Use to judge documentation coverage for every ready requirement in the readiness manifest — ranks synced corpus documents per requirement and gets a covered/partial/missing verdict with verbatim evidence quotes, gaps, and recommendations via the Synapse LLM. MUST run bundled lib/assess.py — never regenerate the assessor. Evidence comes only from /workspace/local/iso-readiness/corpus. No cortex_search.