mindkeg-mcp 0.4.0 → 0.5.0

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
@@ -32,6 +32,7 @@ Unlike traditional RAG systems that chunk large documents, Mind Keg stores **pre
32
32
  - API key authentication with per-repository access control
33
33
  - SQLite storage (zero dependencies, zero config)
34
34
  - Import/export for backup and migration
35
+ - **Smarter knowledge management**: auto-categorization (KNN voting), conflict detection, smart staleness scoring, access tracking with relevance decay, near-duplicate merging, typed learning relationships
35
36
  - **Enterprise security**: encryption at rest, audit logging, TTL/data retention, Prometheus monitoring, rate limiting, content integrity verification
36
37
 
37
38
  ## Quick Start
@@ -171,6 +172,8 @@ Copy `templates/AGENTS.md` to the root of any repository where you want agents t
171
172
  | `deprecate_learning` | Mark a learning as deprecated |
172
173
  | `flag_stale` | Flag a learning as potentially outdated |
173
174
  | `delete_learning` | Permanently delete a learning |
175
+ | `merge_learnings` | Merge near-duplicate learnings into a canonical entry |
176
+ | `relate_learnings` | Create typed relationships between learnings |
174
177
  | `list_repositories` | List all repositories with learning counts |
175
178
  | `list_workspaces` | List all workspaces with learning counts |
176
179
 
@@ -258,7 +261,7 @@ Disables semantic search and falls back to SQLite FTS5 full-text search — all
258
261
 
259
262
  ## Enterprise Security
260
263
 
261
- Mind Keg 0.4.0 ships a suite of security features suitable for corporate and regulated environments.
264
+ Mind Keg ships a suite of security features suitable for corporate and regulated environments.
262
265
 
263
266
  ### Encryption at Rest
264
267
 
@@ -379,6 +382,9 @@ Each learning contains:
379
382
  | `ttl_days` | integer or null | Per-learning TTL; overrides global `MINDKEG_DEFAULT_TTL_DAYS` |
380
383
  | `source_agent` | string or null | Agent name for provenance tracking |
381
384
  | `integrity_hash` | string or null | SHA-256 hash of canonical fields for tamper detection |
385
+ | `access_count` | integer | Times returned by search/get_context (feeds ranking) |
386
+ | `last_accessed_at`| ISO 8601 or null | Last time returned by search/get_context |
387
+ | `staleness_score` | float 0.0–1.0 | Auto-computed from age, access recency, and conflicts |
382
388
  | `created_at` | ISO 8601 | Auto-set on creation |
383
389
  | `updated_at` | ISO 8601 | Auto-updated on modification; TTL expiry anchors to this |
384
390
 
@@ -457,8 +463,8 @@ src/
457
463
  crypto/ AES-256-GCM field encryption
458
464
  monitoring/ Prometheus metrics + /health endpoint
459
465
  security/ Content sanitization, integrity hashing, rate limiter
460
- tools/ One file per MCP tool (9 tools) + shared tool-utils
461
- services/ LearningService + EmbeddingService + PurgeService
466
+ tools/ One file per MCP tool (11 tools) + shared tool-utils
467
+ services/ LearningService + EmbeddingService + PurgeService + ConflictDetector + StalenessEngine
462
468
  storage/ StorageAdapter interface + SQLite impl
463
469
  models/ Zod schemas + TypeScript types
464
470
  utils/ Logger (pino → stderr) + error classes