You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
This commit was created on GitHub.com and signed with GitHub’s verified signature.
Fixed
OpenAI-compatible dispatch with a configured base_url/api_key_env no longer rejects itself (#659). _validate_openai_compatible_agent_command() now accepts the --base-url and --api-key-env flags dispatch.sh emits for env-configured providers, with strict format validation (a non-empty HTTP(S) host, a safe environment-variable name) so the two functions stay in sync.
Concurrent same-second spawns no longer share a task_id and interleave provider output (#661). spawn_agent() and spawn_agent_capture_pid() derived an unsupplied task_id from date +%s alone, so two spawns starting in the same second collided and wrote to the same temp files, silently attributing one provider's answer to another's result. Default task_id generation now uses an OS-guaranteed-unique mktemp reservation.
Gemini dispatch now preserves CLI dotenv fallback and custom CA certificates (#660). The isolated provider environment omits empty API-key variables so gemini-cli can load ~/.gemini/.env, while forwarding non-empty GOOGLE_GEMINI_BASE_URL and NODE_EXTRA_CA_CERTS values needed by relays and custom trust chains.