Repository navigation
docs: define model information management architecture - #10033
Conversation
Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com>
# Conflicts: # scripts/ci/smoke-claude-workflow.test.ts # scripts/ci/smoke-cloud-hypervisor-workflow.test.ts # scripts/ci/smoke-codex-workflow.test.ts
Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com> Copilot-Session: 05030fb2-8413-41a3-9790-349049a61b19
Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com> Copilot-Session: 05030fb2-8413-41a3-9790-349049a61b19
Documentation PreviewDocumentation has been built for this PR. To view locally:
Built from commit c2e7545 |
There was a problem hiding this comment.
🟡 Changes recommended
The proposal needs explicit accounting safeguards and compatible, reproducible migration requirements.
4 open findings
What changed in this PR
Defines a unified model-information subsystem for AWF’s host and API proxy, with implementation deferred to incremental follow-ups.
Changes:
- Specifies provider-qualified identity, pricing, reasoning, compatibility, and provenance.
- Defines generated projections, exports, and updater governance.
- Establishes characterization tests and parity-gated migration.
| File | Description |
|---|---|
| docs/adr/0004-model-information-management.md | Adds the model-information architecture and rollout plan. |
🧠 Review effort: Max
💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.
| Promotional pricing is time-bounded data. It MUST include an expiry or effective | ||
| end and, where known, the successor tariff. A source-code reminder is not an | ||
| acceptable expiry mechanism. |
| - public documentation mapping; | ||
| - static catalogue export; | ||
| - generated documentation tables; and | ||
| - manifests containing schema version, catalogue revision, and generation time. |
| Overlays are validated against a versioned overlay schema at startup. Unknown | ||
| fields and invalid prices fail before requests are accepted. Operator overlays | ||
| have highest configured authority within their declared provider, channel, and | ||
| deployment scope. |
| Add deterministic compilation, migrate maintained facts, and represent the | ||
| current pricing and reasoning gaps as canonical records. Generate existing | ||
| legacy formats without requiring consumers to change. |
|
Looks good. I'd like gh-aw-router-training to eventually sync its pricing and effort data from
|
|
+1 to Ryan's points. Adding the gh-aw side, since gh-aw is the other big consumer of model facts and the ADR doesn't mention it yet:
|
|
@copilot address review and comment feedback |
Co-authored-by: lpcox <15877973+lpcox@users.noreply.github.com>
Updated ADR 0004 to address the review and actionable feedback; committed as |
Define workflow-configured model overlays, configurable central pricing refresh, and model-to-harness version constraints. Clarify tariff component scope and directed protocol compatibility, and wire the requirements into rollout and validation gates. Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com>
|
✅ Copilot review passed with no inline comments. @lpcox Add the |
Assume the default pricing feed is blocked. Define workflow and AWF config source replacement, internal mirrors and local snapshots, no implicit network access, and approved fallback behavior. Extend validation and rollout requirements. Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com>
Use governed provider-qualified models.dev snapshots for inference providers without native AWF pricing support. Preserve precedence, provenance, and offline operation, and cover the fallback in validation and rollout requirements. Co-authored-by: Copilot App <223556219+Copilot@users.noreply.github.com>


Summary
Validation
npx markdownlint --config .markdownlint.json docs/adr/0004-model-information-management.mdgit diff --check -- docs/adr/0004-model-information-management.mdNotes
This PR contains only the ADR. The implementation remains intentionally split into incremental, parity-gated follow-up changes.