Skip to content

fix(auxiliary): preserve custom provider headers - #77281

Closed
tcconnally wants to merge 1 commit into
NousResearch:mainfrom
Perseus-Computing-LLC:fix/upstream-auxiliary-provider-extra-headers
Closed

fix(auxiliary): preserve custom provider headers#77281
tcconnally wants to merge 1 commit into
NousResearch:mainfrom
Perseus-Computing-LLC:fix/upstream-auxiliary-provider-extra-headers

Conversation

@tcconnally

Copy link
Copy Markdown
Contributor

Summary

  • Preserve providers.<name>.extra_headers when Hermes constructs auxiliary OpenAI-compatible clients.
  • Ensures named custom providers retain Cloudflare Access service-auth headers for automatic auxiliary tasks such as compression.
  • Keeps the existing secure pair-header design and does not change provider selection or Ollama binding.

Verification

  • Upstream-main auxiliary/provider tests: 55 passed.
  • Canonical per-file runner: 35/35 passed.
  • Ruff: clean.
  • Live temporary-config acceptance through Cloudflare Access → Pop!_OS → Ollama: UPSTREAM_BASE_AUX_LIVE_OK.

Deployment note

After this upstream fix is included in the managed Hermes image, the existing llama provider can be assigned to auxiliary compression and verified end to end.

@alt-glitch alt-glitch added type/bug Something isn't working P3 Low — cosmetic, nice to have comp/agent Core agent runtime: loop, agent_init, prompt builder, context-compression, responses endpoint area/config Config system, migrations, profiles provider/openai OpenAI / Codex Responses API duplicate This issue or pull request already exists labels Aug 3, 2026
@alt-glitch

Copy link
Copy Markdown
Collaborator

This was generated by AI during triage.

Duplicate of #61346. Both repair named custom-provider extra_headers on auxiliary clients; #61346 is the broader open patch, covering async reconstruction, fallback paths, precedence, and regression coverage.

@tcconnally

Copy link
Copy Markdown
Contributor Author

Closing this as a duplicate of #61346, per the triage note. #61346 is the broader open patch covering named custom-provider extra_headers on auxiliary clients (including async reconstruction). Happy to fold any missing coverage into #61346 if helpful.

@tcconnally tcconnally closed this Aug 3, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

area/config Config system, migrations, profiles comp/agent Core agent runtime: loop, agent_init, prompt builder, context-compression, responses endpoint duplicate This issue or pull request already exists P3 Low — cosmetic, nice to have provider/openai OpenAI / Codex Responses API type/bug Something isn't working

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants