Query CLI: make Know callable, not just browsable

[MW8-ai]

Roadmap item — 'CloudIntelMatrix query CLI so Know is callable, not just browsable.' The same verified facts the site renders, now answerable from a shell or a CI job.

Commands

cim list --category "AI / ML"          # what's tracked
cim show "Generative AI"               # one capability, every provider
cim providers "Object Storage"         # provider comparison
cim gov --capability "Generative AI" --level Partial   # where can it legally live?
cim ai --native                         # AI-native capabilities

(python scripts/cim_query.py …, or npm run query -- ….)

The gate

gov filters by government availability (Full > Partial > Limited > None > Unknown) and exits non-zero when nothing meets the level — so it's also a policy gate. Drop it in CI to fail a build that assumes a capability its target environment doesn't offer:

npm run query -- gov --capability "Kubernetes" --level Limited --provider azure

This is the callable form of the Method's first question — where is this workload even allowed to live? — that the METHOD.md walkthrough opens with.

Fit

• Zero dependencies beyond the standard library; reads the existing data/matrix.json, no schema change.
• Follows the repo's scripts/+verify.py convention (argparse, ROOT/DATA, utf-8 stdout).
• --json on every command for the Method's adapters and downstream tooling.
• 15-check test (scripts/test_cim_query.py, plain-python like verify.py) wired into the verify workflow; query / test:query npm aliases added.

Verified

Data contract still passes (verify.py: 0 errors); test suite 15/15; documented examples produce real output; exit-code contract confirmed (0 on match, 1 on no-match, 2 on bad input).

🤖 Generated with Claude Code