@herjarsa/omo-meta-governor 0.49.0 → 0.50.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
package/README.md CHANGED
@@ -53,6 +53,19 @@ Add as a plugin in your OpenCode config (`~/.config/opencode/opencode.jsonc`):
53
53
  }
54
54
  ```
55
55
 
56
+ ### OpenCode V2 compat (v0.50.0+, dual V1+V2)
57
+
58
+ One package serves both runtimes. V1 (`1.18.x`, `plugin` key) uses the classic
59
+ function entrypoint; V2 (OpenChamber v2 / opencode v2, `plugins` key) uses the
60
+ bridged `setup()` (same scoring, decisions, tools, session-start protocol).
61
+ No config change needed — `omo-meta-governor.jsonc` is read identically, and the
62
+ V1 `plugin` config key is normalized automatically by V2.
63
+
64
+ V2 limitations (no V2 hook equivalent per the official migration guide, logged
65
+ at setup, V1-only for now): `command.execute.before` command-filter,
66
+ `experimental.provider.small_model` override, `experimental.compaction.autocontinue`
67
+ overflow loop guard.
68
+
56
69
  ### Plugin configuration (`omo-meta-governor.jsonc`)
57
70
 
58
71
  The plugin reads its detailed configuration from `omo-meta-governor.jsonc`
@@ -602,6 +615,9 @@ All configuration lives under the `meta_governor` key in
602
615
  |-------|------|---------|-------------|
603
616
  | `saveDecisions` | boolean | `true` | Whether to save decision records. |
604
617
  | `saveLessons` | boolean | `true` | Whether to save lessons. |
618
+ | `autoRemember.enabled` | boolean | `true` | v0.49.1: kill-switch for the Phase 4 auto-remember prompt (warn/escalate/stop → agentmemory_memory_save). Set `false` to disable it entirely. |
619
+ | `autoRemember.cooldownMs` | number | `300000` | v0.49.1: min ms between two auto-remember fires in the same session. |
620
+ | `autoRemember.dedupe` | boolean | `true` | v0.49.1: identical content is never re-queued in the same session. |
605
621
 
606
622
  ### `modelOverride`
607
623
 
@@ -643,9 +659,9 @@ All configuration lives under the `meta_governor` key in
643
659
 
644
660
  | Field | Type | Default | Description |
645
661
  |-------|------|---------|-------------|
646
- | `enabled` | boolean | `false` | Master switch. |
662
+ | `enabled` | boolean | `true` | Master switch (v0.33.1+: defaults on). |
647
663
  | `trigger` | enum | `'firstImplement'` | `'sessionStart'` (first transform) or `'firstImplement'` (once write-like tool observed). |
648
- | `router` | enum | `'both'` | `'aas'` \| `'superpowers'` \| `'both'`. |
664
+ | `router` | enum | `'registry'` | `'registry'` (skill-hub, standalone) \| `'superpowers'` \| `'both'`. v0.50.0: `"both"`/`"superpowers"` delegate via `task(load_skills=[...])` and require the superpowers plugin installed separately — without it only the registry half (`omo_skill_find/get/add`) operates. `'aas'` is deprecated, aliased to `'registry'`. |
649
665
 
650
666
  ### `graphSync`
651
667
 
package/dist/config.d.ts CHANGED
@@ -58,6 +58,12 @@ export interface MetaGovernorPluginConfig {
58
58
  maxLessonsPerSession?: number;
59
59
  saveDecisions?: boolean;
60
60
  saveLessons?: boolean;
61
+ /** v0.49.1: anti-loop guard for the Phase 4 auto-remember prompt. */
62
+ autoRemember?: {
63
+ enabled?: boolean;
64
+ cooldownMs?: number;
65
+ dedupe?: boolean;
66
+ };
61
67
  };
62
68
  /** Model override for MetaGovernor internal LLM usage. */
63
69
  modelOverride?: ModelOverrideConfig;