# io.github.kael-bit/engram (npm · engram-rs-mcp)

Hierarchical memory for AI agents. Three-layer (buffer/working/core) with decay and promotion.

- Trust score: 69/100 (medium)
- Change this week: +23
- Registry status: active
- Liveness: live
- Owner verified: no
- Last scored: 2026-08-03

## Components

- npm · `engram-rs-mcp`: 69/100 (this document), [markdown](https://verifymcp.io/servers/kael-bit-engram/engram-rs-mcp.md), [page](https://verifymcp.io/servers/kael-bit-engram/engram-rs-mcp)

## Channel facts

- Registry: `npm`
- Package: `engram-rs-mcp`
- Version: `0.10.0`
- Transport: `stdio`

## Trust breakdown

How this component scores in each security and reliability category. Every signal is checked automatically from public evidence about the published package, including repeated runs of it in an isolated sandbox, and we only credit what we can confirm. Scores are 0–100 per category. Scoring method: https://verifymcp.io/docs/scoring (what has changed: https://verifymcp.io/docs/scoring/changelog)

Scored 2026-08-03.

- **Supply Chain Security**: 87/100
  - No malware found by supply-chain analysis.
  - Only part of the dependency tree could be resolved (95 of 99), so this covers what we could see, not the whole tree.
  - No install/post-install scripts declared.
  - Only part of the dependency tree could be resolved (95 of 99), so this covers what we could see, not the whole tree.
- **Provenance & Transparency**: 45/100
  - Source repository is publicly reachable at the declared URL.
  - Provenance check failed: no build-provenance attestation is published.
  - Clear OSI-approved license (MIT).
  - Actively maintained (last published 156 days ago).
  - Disclosure check failed: no security disclosure policy was found in the source repository.
- **Schema Quality & AI Usability**: 77/100
  - AI-judged instruction clarity (excellent).
  - Tool/resource definitions use about 1393 tokens (~92/item across 15 items; 15 tools + 0 resources), lean.
  - Usage-examples check failed: none of the tools include examples.
- **Stability & Change Management**: 23/100
  - Stability observed for 7 of 30 days with no destabilising changes; credit accrues until the full window elapses.
- **Tool Coverage**: 100/100
  - 100% of tools have a non-trivial description (not blank, and not just the tool's name).
  - 100% of tool parameters carry a description.
- **Capabilities**: 100/100
  - Implements a supported MCP spec version (2025-11-25); the latest is 2026-07-28.

## Install

### Claude

```bash
claude mcp add kael-bit-engram -- npx -y engram-rs-mcp
```

### Codex

```bash
codex mcp add kael-bit-engram -- npx -y engram-rs-mcp
```

### opencode

```json
{
  "$schema": "https://opencode.ai/config.json",
  "mcp": {
    "kael-bit-engram": {
      "type": "local",
      "command": [
        "npx",
        "-y",
        "engram-rs-mcp"
      ],
      "enabled": true
    }
  }
}
```

### OpenClaw

```bash
openclaw mcp add kael-bit-engram --command npx --arg -y --arg engram-rs-mcp
```

### Hermes

```yaml
mcp_servers:
  kael-bit-engram:
    command: "npx"
    args: ["-y", "engram-rs-mcp"]
```

### Other

```json
{
  "mcpServers": {
    "kael-bit-engram": {
      "command": "npx",
      "args": [
        "-y",
        "engram-rs-mcp"
      ]
    }
  }
}
```

## Changelog

Every change recorded for this component, newest first. Days that predate change tracking, or that we cannot explain, say so: "we were watching and nothing happened" and "we were not watching" are different claims.

### 2026-08-02 (score 69, +48)

- [security regression] Provenance: unverified → fail
- [security improvement] Known CVEs: unverified → partial
- [security improvement] Malware scan: unverified → pass
- [security improvement] Install scripts: unverified → pass
- [functional improvement] MCP protocol: unverified → pass
- [functional improvement] Stability: unverified → 0.20
- [functional improvement] Schema quality: unverified → excellent
- [functional improvement] Dependency health: unverified → partial
- [functional improvement] Maintenance: unverified → pass
- [functional improvement] License: unverified → pass
- [functional] Licence: MIT

### 2026-07-31 (score 21, −7)

- [functional] We updated how we score, so this day's move reflects our rubric, not a change to the server

### 2026-07-30 (score 28, −18)

- [security regression] Malware scan: pass → unverified

### 2026-07-27 (score 46)

First indexed and scored.

## MCP tools (15)

### `engram_store` (~222 tokens)

Store a memory. All memories start in Buffer and promote to Working/Core through access frequency and LLM quality gating. Procedural memories and lessons (tag=lesson) auto-promote to Working after 2h. Use supersedes to replace outdated memories by their ids.

Input parameters:

- `content` (string, required): Memory content text
- `importance` (number): Importance 0-1 (default 0.5, bumps +0.02 on each recall)
- `kind` (string): Memory type: semantic (facts, default), episodic (events), procedural (how-to, persists indefinitely)
- `namespace` (string): Override namespace. Use "default" for cross-project knowledge (user identity, preferences, universal lessons)
- `skip_dedup` (boolean): Skip near-duplicate detection
- `source` (string): Source identifier
- `supersedes` (array): IDs of old memories this one replaces
- `sync_embed` (boolean): Wait for embedding generation before returning (default: false)
- `tags` (array): Tags for categorization

### `engram_recall` (~291 tokens)

Hybrid semantic + keyword search with budget-aware retrieval. Fast by default (~30ms cached, ~1s first query). Optional rerank/expand add LLM calls (+2-4s each) — only use when needed.

Input parameters:

- `budget_tokens` (integer): Max token budget
- `dry` (boolean): Skip touch/reinforcement on results (for background queries)
- `expand` (boolean): LLM query expansion (+1-2s). Use for short/vague queries like single words
- `layers` (array): Filter by layers
- `limit` (integer): Max results
- `min_importance` (number): Min importance threshold
- `min_score` (number): Drop results below this score (0-1)
- `namespace` (string): Filter by namespace (also includes "default" namespace automatically)
- `query` (string, required): Search query
- `rerank` (boolean): Re-rank via LLM (+2-4s). Use only when result ordering is critical
- `since` (integer): Only memories created after this unix ms timestamp
- `sort_by` (string): Sort order (default: score)
- `source` (string): Filter by source (e.g. session, extract, api)
- `tags` (array): Filter by tags (must have ALL specified)
- `until` (integer): Only memories created before this unix ms timestamp

### `engram_recent` (~105 tokens)

List recent memories by creation time. Good for session context recovery.

Input parameters:

- `hours` (number): Look back N hours (default 2)
- `layer` (integer): Filter by layer
- `limit` (integer): Max results (default 20)
- `min_importance` (number): Minimum importance threshold
- `namespace` (string): Filter by namespace (also includes "default" namespace automatically)
- `source` (string): Filter by source (e.g. session)

### `engram_resume` (~182 tokens)

Full memory bootstrap for session recovery. Returns core (permanent knowledge), working (ongoing context/decisions), buffer (transient), recent activity, and session notes. Use workspace tags to filter by current work context. Compact mode (default) minimizes token usage.

Input parameters:

- `budget` (number): Max chars across all sections (default 8000, 0=unlimited)
- `compact` (boolean): Compact output: content+tags only (default true)
- `hours` (number): Look back N hours for recent/sessions (default 4)
- `limit` (number): Max memories per layer (default 100)
- `namespace` (string): Filter by namespace (also includes "default" namespace automatically)
- `workspace` (string): Comma-separated workspace tags to filter core/working memories (e.g. 'engram,rust')

### `engram_extract` (~55 tokens)

Extract structured memories from raw text using LLM. Feed conversation logs or notes and get individual memories.

Input parameters:

- `auto_embed` (boolean): Generate embeddings (default true)
- `text` (string, required): Raw text to extract memories from

### `engram_search` (~64 tokens)

Quick keyword search. Lighter than recall — no scoring or budget logic.

Input parameters:

- `limit` (integer): Max results (default 10)
- `namespace` (string): Filter by namespace (also includes "default" namespace automatically)
- `q` (string, required): Search query

### `engram_consolidate` (~111 tokens)

Run a memory consolidation cycle. Promotes important memories upward, drops decayed entries. With merge=true, uses LLM to merge similar memories.

Input parameters:

- `decay_drop_threshold` (number): Drop below this recency (default 0.01)
- `merge` (boolean): Use LLM to merge similar memories
- `promote_min_importance` (number): Min importance to promote (default 0.6)
- `promote_threshold` (integer): Min access count to promote (default 3)

### `engram_stats` (~22 tokens)

Get memory statistics: counts per layer, AI status, version.

### `engram_repair` (~35 tokens)

Repair FTS search index. Removes orphaned entries and rebuilds missing ones. Safe to run anytime — idempotent.

### `engram_health` (~26 tokens)

Detailed health check: uptime, RSS memory, embed cache stats, AI config status.

### `engram_triggers` (~61 tokens)

Fetch trigger memories for a specific action. Call before performing an action (e.g. git-push, deploy) to recall relevant lessons and rules.

Input parameters:

- `action` (string, required): Action name (e.g. git-push, deploy, send-message)

### `engram_delete` (~38 tokens)

Delete a memory by ID. Use when a memory is outdated, incorrect, or redundant.

Input parameters:

- `id` (string, required): Memory ID to delete

### `engram_update` (~109 tokens)

Update a memory's content, metadata (importance, tags, layer, kind). Use supersedes on store to replace with history.

Input parameters:

- `content` (string): New content text
- `id` (string, required): Memory ID to update
- `importance` (number): New importance value
- `kind` (string): Memory type
- `layer` (integer): Move to layer: 1=buffer, 2=working, 3=core
- `tags` (array): Replace tags

### `engram_trash` (~36 tokens)

List soft-deleted memories. Recently deleted items can be restored.

Input parameters:

- `limit` (integer): Max results (default 100)

### `engram_restore` (~36 tokens)

Restore a soft-deleted memory from trash back to its original layer.

Input parameters:

- `id` (string, required): Memory ID to restore from trash

## Diagnostics

Captured diagnostic sections: Provenance, Dependencies. The full working is on the page: https://verifymcp.io/servers/kael-bit-engram/engram-rs-mcp#diagnostics

## Score history

- 2026-08-03: 69
- 2026-08-02: 69
- 2026-08-01: 21
- 2026-07-31: 21
- 2026-07-30: 28
- 2026-07-28: 46
- 2026-07-27: 46

## Links

- npm package: https://www.npmjs.com/package/engram-rs-mcp
- Socket report: https://socket.dev/npm/package/engram-rs-mcp
- Repository: https://github.com/kael-bit/engram-rs
- Changelog RSS feed: https://verifymcp.io/servers/kael-bit-engram/engram-rs-mcp/changelog.xml
- Changelog JSON feed: https://verifymcp.io/servers/kael-bit-engram/engram-rs-mcp/changelog.json
- HTML version of this page: https://verifymcp.io/servers/kael-bit-engram/engram-rs-mcp