[P2.19] Provider-selection gaps: pre-flight key check and Think-phase ignore `--llm-provider`

[frankbria]

Severity: medium | Category: correctness | Phase: P2.19

Problem

work start/batch/retry validate ANTHROPIC_API_KEY based on engine only, ignoring --llm-provider/CODEFRAME_LLM_PROVIDER/config — so built-in-engine + OpenAI/Ollama workflows are blocked without an Anthropic key and OPENAI_API_KEY is never checked. prd stress-test and tasks generate are hardcoded to Anthropic, so an OpenAI-only user cannot run the Think phase at all.

Evidence

codeframe/cli/app.py:2433 (key check), :1712 (stress-test), :1859 (tasks generate)

Acceptance criteria

A shared helper resolves the effective provider (flag→env→config→default) and validates the matching key; Think-phase commands route through the same provider resolution or explicitly document Anthropic-only.

Dependencies

None

---

_{Filed from the SaaS launch-readiness audit. Atomic: one developer, one session. Work order: strictly P0.1 → P3.12 (no forward dependencies).}

[coderabbitai]

🔗 Related PRs

#383 - fix: fail loudly when ANTHROPIC_API_KEY is missing (#376) [merged]
#551 - [4.D.2] Provider factory + env var wiring (get_provider, ModelSelector) [merged]
#552 - feat: multi-provider LLM support (#545–549) [merged]
#588 - feat(settings): API key management tab (#555) [merged]

---

📝 Issue Planner

_{Check the box below or use the @coderabbitai plan command to generate an implementation plan and prompts that you can use with your favorite coding assistant.}

• Create Plan

---

🧪 Issue enrichment is currently in open beta.

To disable automatic issue enrichment, add the following to your .coderabbit.yaml:

issue_enrichment:
  auto_enrich:
    enabled: false

💬 Have feedback or questions? Drop into our discord!