feat(llm): ddcdragon provider — makes CEF QA actually run on the new layout - #7
Open
jaxter03 wants to merge 1 commit into
Open
feat(llm): ddcdragon provider — makes CEF QA actually run on the new layout#7jaxter03 wants to merge 1 commit into
jaxter03 wants to merge 1 commit into
Conversation
The CEF hiring-coach QA runs its investigation on the DDC inference gateway
(LLM_PROVIDER=ddcdragon, OSS gemma4_31b, keyless). That provider existed on the
old branch but was never ported to the new layout, so opensre cef-qa failed with
'Unsupported LLM provider ddcdragon'.
Adds it per core/llm/AGENTS.md (custom SDK client — bespoke {model,input} ->
{output:{text,tool_calls}} envelope, not OpenAI-wire/LiteLLM):
- DdcDragonLLMClient / DdcDragonAgentClient in core/llm/transports/sdk/*
- construction branches in client_builders.py
- config literal + DDCDRAGON_* settings + ProviderSpec (credential_kind=ambient,
keyless) + wizard option + env_sync non-secret keys
- docs/llm-providers.mdx row; tests/core/runtime/llm/test_ddcdragon_provider.py
Validated live: the ported client returns 'OK' from the gateway; get_llm(AGENT)
resolves to DdcDragonAgentClient; opensre cef-qa runs the full investigation
end-to-end on gemma4_31b (grounded verdict + report).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PkNM3j2VZmxijQsWFZMPiy
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Companion to the CEF agent-QA port (#5).
opensre cef-qaruns its investigation on the DDC inference gateway (LLM_PROVIDER=ddcdragon, keyless OSSgemma4_31b). That provider existed on the old branch but was never ported — so on the new layoutcef-qadied withUnsupported LLM provider 'ddcdragon'. This live-caught gap is what a real end-to-end run surfaced.What lands
Per
core/llm/AGENTS.md"Adding a Hosted API Provider" (custom SDK client — bespoke{model:{bucket,name,version}, input}→{output:{text,tool_calls}}envelope, not OpenAI-wire/LiteLLM):DdcDragonLLMClient/DdcDragonAgentClientincore/llm/transports/sdk/*(override transport, reuse OpenAI message/tool machinery)core/llm/client_builders.pyconfig/config.pyliteral +DDCDRAGON_*settings +DDCDRAGON_LLM_CONFIG;ProviderSpec(credential_kind=ambient, keyless) + wizardProviderOption+env_syncnon-secret keysdocs/llm-providers.mdxrow;tests/core/runtime/llm/test_ddcdragon_provider.pyValidation (live, not mocked)
make typecheckclean (1489 files); 7 provider unit tests + 107 regression (builders/factory/wizard/catalog) pass; ruff clean.OK.get_llm(AGENT)withLLM_PROVIDER=ddcdragonresolves toDdcDragonAgentClient(gemma4_31b), no error.opensre cef-qaend-to-end on the new layout: drives ddcdragon LLM → CEF tools → vault/grafana → grounded verdict + report. (Verifiedno_gocorrectly diagnosed stale test creds — the QA detecting a real failure, exactly as designed.)Base = the CEF-port branch (#5) so the diff is ddcdragon-only; retarget to
mainonce #5 merges. Together, #5 + this make CEF QA functional onmain.🤖 Generated with Claude Code