docs(where-we-stand): a scan without a provider key records learning_endpoint_unavailable

cli.py appends the degradation when resolve_chat_endpoint returns nothing; the
gap is that report.md has no line for it, not that nothing is recorded.

Co-Authored-By: Claude Fable 5.1 <[email protected]>
Claude-Session: https://claude.ai/code/session_01EmM1j46qgMh3zqD1QMwEo
This commit is contained in:
Marius Ciepluch
2026-10-02 19:55:22 +02:00
co-authored by Claude Fable 5.1
parent 6bd44caa50
commit 515f94e3c2
+5 -6
View File
@@ -250,9 +250,8 @@ Infrastructure.
Instrument gaps. Instrument gaps.
- A missing provider key is a silent path in a scan. The plane refuses to start a Task whose Model - A missing provider key is a quiet path in a scan. The plane refuses to start a Task whose Model
names an unset variable (`plane/router.py`, `StartError`), but a scan without a key asks the names an unset variable (`plane/router.py`, `StartError`). A scan without a key asks the model
model nothing, reports only the graph's entailed findings, and records no degradation for the nothing, reports only the graph's entailed findings, and records `learning_endpoint_unavailable`
skipped model layer (`cli.py` writes degradations for `hunter_model_unavailable`, in `manifest.json` (`cli.py`, after `resolve_chat_endpoint` returns nothing); `report.md` has no
`cpg_unavailable` and `learning_endpoint_unavailable` only). The rule that a scan which could line for that reason yet, so the report does not say the model layer was skipped.
not look must say so does not yet cover this case.