docs(specs): cgis init-ontology design — auto-proposed patterns.yaml (#174)

[zaebee]

Summary

Design spec for #174 (roadmap #179 P0, the adoption-blocker half of the pair with merged #178): cgis init-ontology proposes a ready-to-edit patterns.yaml from the measured graph — measure-then-label instead of a blank page.

Key decisions:

• DriftScorer IS the fitter (approach A): per-domain fit = score() against each of the 5 bundled templates via a synthetic DomainConfig — zero new fitting math; layer weights / empty-layer exclusion / unresolved discount inherited from _score_v2, so proposed numbers stay comparable to later drift runs
• Honest no-fit branch (user-decided): best fit > 0.5 → hygiene-only + # no template fits (best: X at 0.55) — no forced labels (the httpx mislabel trap from 🗺️ Roadmap: priorities from multi-repo dogfooding (#170–#178) #179); below min_nodes / zero edges → hygiene-only with reasons. Hygiene-only entries also get measured+margin tolerance so the round-trip guarantee covers cycle-bearing unlabeled domains
• Round-trip acceptance: propose → analyze_drift on the same graph → no empty, nothing above its proposed tolerance, gate green by construction; on fixtures and the cgis self-graph
• Surfaces (user-decided): CLI init-ontology (overwrite-refusal without --force) + MCP cgis_init_ontology returning YAML text read-only — no file writes from MCP (avoids the S2083 taint class)
• Bundled default header (profiles/patterns/hygiene, no cgis domains) as a module constant, pinned against docs/ontology/patterns.yaml by a staleness test
• Lane A discipline: new query/ontology_init.py module, append-only CLI/MCP blocks (parallel-session contract)

Test Plan

• Spec review (gemini + guardian + colleague — their onboarding pain)
• Implementation PR follows (plan → SDD); it closes Add cgis init-ontology: auto-propose patterns.yaml from measured graph #174

🤖 Generated with Claude Code

[zaebee]

/gemini review

[zaebee]

/guardian review

[github-actions]

LGTM — no defects found in this diff.

Diff is a design document (spec) with no executable code changes. No logic, type, contract, ontology, or test coverage defects to report.

---

🤖 mistral-medium-latest · 5,734 prompt + 41 completion = 5,775 tokens

[gemini-code-assist]

Code Review

This pull request introduces a design specification for cgis init-ontology to auto-propose a patterns.yaml file. The feedback suggests explicitly validating that the SQLite database file exists before opening a connection in both the propose_ontology flow and the MCP tool to prevent SQLite from silently creating empty database files.

Important

The consumer version of Gemini Code Assist on GitHub is being sunset. Starting June 18, 2026, new organization installations will be blocked, and all code review activity will officially cease on July 17, 2026.
For more details on the timeline and next steps, please review the Help Documentation.

[gemini-code-assist]

Code Review

This pull request introduces a design specification for the cgis init-ontology feature, which automatically proposes a patterns.yaml configuration to simplify onboarding. The feedback suggests updating the proposed flow to explicitly validate the existence of the SQLite database file before opening the store, preventing SQLite from silently creating an empty database file.

Important

The consumer version of Gemini Code Assist on GitHub is being sunset. Starting June 18, 2026, new organization installations will be blocked, and all code review activity will officially cease on July 17, 2026.
For more details on the timeline and next steps, please review the Help Documentation.

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud