mirror of
https://github.com/NousResearch/hermes-agent.git
synced 2026-04-25 00:51:20 +00:00
ab4ba8163a
* feat(migration): comprehensive OpenClaw -> Hermes migration v2 Extends the existing migration script from ~15% to ~95% coverage of OpenClaw's configuration surface. Adds 17 new migration modules: Direct migrations (written to config.yaml/.env): - MCP servers: full server definitions with transport, tools, sampling - Agent defaults: reasoning_effort, compression, human_delay, timezone - Session config: reset triggers (daily/idle) -> session_reset - Full model providers: custom_providers with base_url/api_mode - Deep channel config: Matrix, Mattermost, IRC, Discord deep settings - Browser config: timeout settings - Tools config: exec timeout -> terminal.timeout - Approvals: mode mapping (smart/manual/auto -> Hermes equivalents) Archived for manual review (no direct Hermes equivalent): - Plugins config + installed extensions - Cron jobs (with note to use 'hermes cron') - Hooks/webhooks config - Multi-agent list + routing bindings - Gateway config (port, auth, TLS) - Memory backend config (QMD, vector search) - Skills registry per-entry config - UI/identity settings - Logging/diagnostics preferences Also adds: - MIGRATION_NOTES.md generation with PM2 reassurance message - _set_env_var helper for consistent env file management - Updated presets to include all new options - Comprehensive mock test passing (12 migrated, 12 archived) * feat(migration): add terminal recap with visual summary Replaces raw JSON dump with a formatted box showing migrated/archived/ skipped/conflict/error counts, detailed item lists with labels, PM2 reassurance message, and actionable next steps. JSON output available via MIGRATION_JSON_OUTPUT=1 env var. * fix(test): allowlist python_os_environ as known false-positive in skills guard test MIGRATION_JSON_OUTPUT env var is a legitimate CLI feature flag that enables JSON output mode, not an env dump. Add it alongside agent_config_mod as an accepted finding in test_skill_installs_cleanly_under_skills_guard. * fix(test): add hermes_config_mod to known false-positives in skills guard test The scanner flags two print statements that tell the user to *review* ~/.hermes/config.yaml in the post-migration summary. The script never writes to that file — those are informational strings, not config mutations. --------- Co-authored-by: Hermes <hermes@nousresearch.ai> |
||
|---|---|---|
| .. | ||
| autonomous-ai-agents | ||
| blockchain | ||
| creative | ||
| email/agentmail | ||
| health | ||
| mcp | ||
| migration | ||
| productivity/telephony | ||
| research | ||
| security | ||
| DESCRIPTION.md | ||