Hacktoberfest 2026: le issue che i maintainer hanno segnato per ottobre, aperte e adatte ai principianti. Sfoglia le issue Hacktoberfest

bug(agents): built-in agents cannot inherit an OAuth-backed host LLM

Chiusa
  • Qualcosa non va?
#809 0 commenti 0 reazioni 0 assegnatari Vedi su GitHub

I maintainer di solito rispondono entro 1 giorno

Nessuno ha ancora preso questa issue.

Valutazione

Difficoltà
4/5
Tempo stimato
3-5 giorni
Idoneità per principianti
68/100
Tipo di issue
Bug
Chiarezza
Specificata chiaramente
Stato di attività
Attiva
Stack tecnologico
python
Ambito
ai, authentication

Direzione di ricerca

Start in raven/config/product_render.py at inherit_llm, then trace the main runtime's provider-readiness or credential-resolution logic referenced by the issue. Add coverage for an OpenAI Codex OAuth host with no provider apiKey, while preserving own-key and API-key inheritance behavior. Done means all four built-in products inherit the host provider and resolve its OAuth credentials.

Scritto dal modello di indicizzazione a partire dal testo della issue.

Descrizione

Summary

The four shipped agent products fail to launch when the host Raven uses an OAuth-backed provider such as OpenAI Codex, even though the host model is authenticated and works.

The products are documented to use either their own key or the host Raven LLM. The host-inheritance branch currently treats only a provider block containing an apiKey as usable, so an authenticated OAuth provider is rejected before the child runtime starts.

Environment

  • Raven: 0.2.2
  • Platform: macOS
  • Host provider: openai_codex
  • Host model: an account-catalog Codex model
  • Installation: published Raven tool/wheel

Reproduction

  1. Sign in with raven provider login openai-codex.
  2. Select an OpenAI Codex model as the main Raven model.
  3. Verify that raven doctor --probe succeeds.
  4. Open Agent Connector and re-test Raven-Code, Raven-Design, Raven-Oncall, or Raven-Research.
  5. Observe that each product starts and immediately exits.

The launchers report:

Raven-Code: CODE_API_KEY is not set and the host config has no provider key to inherit from
Raven-Design: configure a model provider in the host Raven settings before starting Design
Raven-Oncall: ONCALL_API_KEY is not set and the host config has no provider key to inherit from
Raven-Research: RESEARCH_API_KEY is not set and the host config has no provider key to inherit from

At the same time:

raven provider list
openai_codex  OpenAI Codex  OAuth  configured

raven doctor --probe
LLM Probe: success

Root cause

raven/config/product_render.py::inherit_llm currently gates inheritance on an API key stored inside a provider block:

providers = host.get("providers") or {}
if not any(isinstance(p, dict) and p.get("apiKey") for p in providers.values()):
    return ""

OAuth credentials are intentionally stored outside the provider block, so a working OAuth provider has no apiKey there. The function therefore returns an empty result before copying the host provider, routing, and model configuration.

This affects all four visible shipped products because their no-own-key paths call inherit_llm directly or depend on the same result.

Expected behavior

A host provider that is authenticated and usable through OAuth should be eligible for built-in-agent inheritance, just like an API-key provider. The child should receive the host provider/model binding and resolve the same OAuth credential store.

Actual behavior

The main Raven can answer through OpenAI Codex OAuth, but every built-in professional agent fails its handshake and cannot be dispatched.

Suggested fix scope

  • Replace the apiKey-presence gate with the same provider-readiness or credential-resolution logic used by the main runtime.
  • Ensure the rendered child process resolves the host OAuth credential store.
  • Add coverage for a host using OpenAI Codex OAuth with no provider apiKey.
  • Keep existing own-key and API-key inheritance behavior unchanged.

Privacy

This report contains no credential values, account identifiers, private paths, or private conversation data.

Lingua principale
Python
Stelle
4.1k
Fork
94
Merge medio
12h 27m
PR unite (30g)
400

Preparare l'ambiente

Questo progetto non fornisce container di sviluppo, Dockerfile né guida per i contributori, quindi l'ambiente è a tuo carico: parti dal suo README e consulta la nostra guida al primo contributo per i passaggi generali.

Come iniziare

  1. Leggi tutta la issue e poi la guida ai contributi del progetto.
  2. Commenta sulla issue per dire che te ne occupi tu — evita che due persone facciano lo stesso lavoro.
  3. Fai un fork del repository e lavora su un branch.
  4. Apri una pull request che faccia riferimento al numero della issue.

Altre issue di EverMind-AI/Raven

Tutte le issue di EverMind-AI/Raven

Issue simili

Altre issue su Python

Ricevi le nuove issue nella tua casella

Un breve riepilogo di issue GitHub adatte ai principianti.