mirror of
https://github.com/NousResearch/hermes-agent.git
synced 2026-07-25 17:18:11 +00:00
040e18ad90
initialize() snapshots OPENVIKING_* into the provider once, so /reload (which only updates os.environ) leaves viking_* tools running against stale auth — users have to restart hermes to pick up keys added to ~/.hermes/.env after startup (#21130). Add _ensure_client(), which re-resolves the connection settings via the same _resolve_connection_settings/_load_hermes_openviking_config path initialize() uses and rebuilds + health-checks the client only when an OPENVIKING_* value actually changed; otherwise it reuses the cached client (or a cached None for a known-down target) so the hot path stays at one dict comparison with no network calls. Route every live client-gated path through it: system_prompt_block, prefetch (pre-turn recall), sync_turn, on_session_end, on_session_switch (session rotation), on_memory_write and handle_tool_call. The unreachable branch calls _handle_runtime_openviking_unreachable() so a newly-resolved but not-yet-up LOCAL endpoint keeps initialize()'s recovery — (re)start the server and attach in the background — instead of silently disabling memory. Refreshing is gated behind a flag set at the end of initialize() so the baseline is established before any env re-resolution happens — callers that wire up a client directly (e.g. tests) keep it untouched. Refs #21130 |
||
|---|---|---|
| .. | ||
| browser | ||
| context_engine | ||
| cron_providers | ||
| dashboard_auth | ||
| disk-cleanup | ||
| google_meet | ||
| hermes-achievements | ||
| image_gen | ||
| kanban | ||
| memory | ||
| model-providers | ||
| observability | ||
| platforms | ||
| security-guidance | ||
| spotify | ||
| teams_pipeline | ||
| video_gen | ||
| web | ||
| __init__.py | ||
| plugin_utils.py | ||