Gate live E2E tests behind explicit provider selection and add OpenAI-compatible live path

Add repeatable OpenAI-compatible provider backend configuration to a TypeScript runtime

Provider abstraction existed but boot path still hard-coded a single LLM provider

Clean-break standalone migration still had dual-mode config defaults and stale tests

Thinking-capable local Ollama agents produced empty scorable output in a [redacted:name] benchmark harness

Sandbox benchmark agents to prevent local answer-key leakage

CTF benchmark over-scored wrong-location findings and leaked answer hints in cold prompts

Silhouette-only k-means split refuses big-soupy clusters; forced-k fallback with cohesion-improvement gate splits the residue

Blender Python: multi-object edit mode silently corrupts other meshes' UVs/normals

PSX vertex snap on world-space UV materials causes texture "boiling" on flat surfaces

React snapshot-on-settle timer captures empty data when streaming flushes are throttled

[redacted:name] Qwen3 benchmark agents emit thinking-only output and schema-mismatched findings

Python on Windows: UnicodeEncodeError 'charmap' codec can't encode '✅' (✅) in print() — script crashes after work succeeds

tectonic 'Undefined control sequence' when markdown code spans contain Greek letters (e.g. θ, α)

[redacted:name] benchmark runner launched duplicate models because per-wave concurrency repeated the wave model

Add local Ollama-backed model trial to a TypeScript benchmark while preserving CLI agent tooling

Reconstruct a branch by removing one author's commits while preserving other authors' changes

CTF benchmark harness used local throwaway agents instead of provided real agent keys

Gemini benchmark agents hit 429 because free-tier daily project/model quota was exhausted, not just RPM burst traffic

React Three.js graph activity pulses kept renderer RAF loop alive indefinitely