# behold (npm · @intentius/behold)

Your infrastructure repo as a graph, from the files. Draws Terraform roots, no state needed.

- Trust score: 77/100 (medium)
- Change this week: +18
- Registry status: active
- Liveness: live
- Owner verified: no
- Last scored: 2026-11-22

## Components

- npm · `@intentius/behold`: 77/100 (this document), [markdown](https://verifymcp.io/servers/intentius-behold/intentius-behold.md), [page](https://verifymcp.io/servers/intentius-behold/intentius-behold)

## Channel facts

- Registry: `npm`
- Package: `@intentius/behold`
- Version: `0.34.4`
- 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 2026-11-22.

- **Supply Chain Security**: 81/100
  - No malware found by supply-chain analysis.
  - CVE check failed: a known high-severity CVE affects @modelcontextprotocol/sdk 1.30.1, a direct dependency. A fixed version is available.
  - No install/post-install scripts declared.
  - 43 of 168 dependencies flagged as unhealthy.
- **Provenance & Transparency**: 97/100
  - Source repository is publicly reachable at the declared URL.
  - Cryptographically verified build provenance (signed, bound to INTENTIUS/behold).
  - Clear OSI-approved license (Apache-2.0).
  - Actively maintained (last published 0 days ago).
  - Disclosure check failed: no security disclosure policy was found in the source repository.
- **Schema Quality & AI Usability**: 67/100
  - 100% of prompts and resources have a non-trivial description (not blank, and not just the item's name).
  - AI-judged instruction clarity (good).
  - Context-footprint check failed: tool/resource definitions use about 3378 tokens (~482/item across 7 items; 6 tools + 1 resources), over budget; trim descriptions and params.
  - 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.
- **Tool Safety**: 100/100
  - No prompt-injection markers were found in the server instructions, tool names or descriptions we captured.
  - We read all 6 captured tool definition(s), and no name or description among them implies an irreversible operation.
  - An AI judge read all 8 captured unit(s) of tool text and found none that tries to manipulate the model reading it.
- **Capabilities**: 100/100
  - Implements a supported MCP spec version (2025-11-25); the latest is 2026-07-28.
  - Supports UI / widget rendering.

## Install

### How do I install the behold MCP server?

behold runs locally as an npm package, launched with npx -y @intentius/behold mcp. Ready-made configuration for Claude, Cursor, VS Code, Codex and 5 more is on this page, copied from each client's own documentation.

### Claude

```bash
claude mcp add intentius-behold -- npx -y @intentius/behold mcp
```

### Cursor

```json
{
  "mcpServers": {
    "intentius-behold": {
      "command": "npx",
      "args": [
        "-y",
        "@intentius/behold",
        "mcp"
      ]
    }
  }
}
```

### VS Code

```json
{
  "servers": {
    "intentius-behold": {
      "command": "npx",
      "args": [
        "-y",
        "@intentius/behold",
        "mcp"
      ]
    }
  }
}
```

### Codex

```bash
codex mcp add intentius-behold -- npx -y @intentius/behold mcp
```

### opencode

```json
{
  "$schema": "https://opencode.ai/config.json",
  "mcp": {
    "intentius-behold": {
      "type": "local",
      "command": [
        "npx",
        "-y",
        "@intentius/behold",
        "mcp"
      ],
      "enabled": true
    }
  }
}
```

### OpenClaw

```bash
openclaw mcp add intentius-behold --command npx --arg -y --arg @intentius/behold --arg mcp
```

### Hermes

```yaml
mcp_servers:
  intentius-behold:
    command: "npx"
    args: ["-y", "@intentius/behold", "mcp"]
```

### Netclaw

```json
{
  "McpServers": {
    "intentius-behold": {
      "Transport": "stdio",
      "Command": "npx",
      "Arguments": [
        "-y",
        "@intentius/behold",
        "mcp"
      ]
    }
  }
}
```

### Vellum

```bash
assistant mcp add intentius-behold -t stdio -c npx -a -y @intentius/behold mcp
```

### Other

```json
{
  "mcpServers": {
    "intentius-behold": {
      "command": "npx",
      "args": [
        "-y",
        "@intentius/behold",
        "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.

### 2026-11-22 (score 77, +15)

- [security improvement] Malware scan: unverified → pass

### 2026-11-21 (score 62, −14)

- [security regression] Malware scan: pass → unverified
- [security regression] Known CVEs: fail → unverified
- [functional regression] Dependency health: 0.90 → unverified
- [functional] Package version: 0.34.2 → 0.34.4
- [functional] Package version: 0.34.2 → 0.34.3

### 2026-11-19 (score 76, +16)

- [security improvement] Malware scan: unverified → pass

### 2026-11-18 (score 60, −15)

- [security regression] Known CVEs: fail → unverified
- [security regression] Malware scan: pass → unverified
- [functional regression] Dependency health: 0.91 → unverified
- [functional] Package version: 0.33.3 → 0.34.2
- [functional] Package version: 0.33.3 → 0.34.1
- [functional] Package version: 0.33.3 → 0.34.0

### 2026-11-17 (score 75, +21)

- [security regression] Known CVEs: unverified → fail
- [security improvement] Malware scan: unverified → pass
- [functional improvement] Dependency health: unverified → 0.91

### 2026-11-16 (score 54, −5)

- [security regression] Tool safety: pass → unverified
- [security regression] Known CVEs: fail → unverified
- [security improvement] Malware scan: unverified → pass
- [security] Stability: Stability not yet verified: we do not have a sandbox capture of the MCP schema this version of the package serves yet.
- [functional regression] Capabilities: pass → unverified
- [functional regression] Tool coverage: 100 → unverified
- [functional regression] Schema quality: 100 → unverified
- [functional regression] Dependency health: 0.90 → unverified
- [functional improvement] Stability: unverified → 0.03
- [functional] Package version: 0.32.0 → 0.33.3
- [functional] Package version: 0.32.0 → 0.33.2
- [functional] Package version: 0.32.0 → 0.33.1
- [functional] Package version: 0.32.0 → 0.33.0

### 2026-11-15 (score 59)

First indexed and scored.

## MCP tools (6)

### `ask` (~1213 tokens)

Ask a chant search query of the estate behold serves, and get chant's answer: its read-contract search document, verbatim, with `link` (the behold view that draws it) and `drawn: false` added. ask never moves the person's tab; use show for that. On a declared workspace the document is chant workspace search's, with every member's rows as <member>/<id>; on one project or a Terraform folder it is chant search --json's. Uses the running behold serve of this project; with none, behold answers in this process. Needs chant 0.124.46 or newer. Which demo questions it answers: "what's in here?" is countBy kind with the query kind:aws_ (or empty); "how many resources does each module own, by type?" is countBy module, then countBy kind with the query narrowed to a module; "what would this change do, and why does it replace the database?" is plan with the query attr:planAction=replace and show planActions,planReplacePaths,planActionReason. Beside a chant search query, ask takes cypher (a read-only Cypher query over every member's graph and every index in one LadybugDB, behold#1027) or resource (resolve a name, id, address or cloud id, then a template). The query is chant's search grammar, unchanged (https://intentius.io/chant/cli/search/). env reads that environment live (chant search --live --env <env>); snapshot with env replays a recorded snapshot instead (chant --snapshot <ref>, such as latest); chant's --explain is always on, for the near-misses drawn. It takes no --check-live or --check-snapshot (drift is not drawn), and no --ambient, --src or --fail-on-drift. Read the capabilities tool first: it lists the capabilities a cap: term asks (cap:internet-reachable(22), ->[cap=network]), each one's meaning, parameters and rule (INTENTIUS/chant#4375).

Input parameters:

- `countBy` (string): Group the matches the way chant search --count-by does: kind, module or attr:<name> (behold#922). The answer is chant's groups and total; the view lights each group's cards and labels the group with…
- `cypher` (string): A read-only Cypher query over the graph pane's database (behold#1027): every member's chant IR, every engine graph document (choudoufu, plunge) and every index version, in one LadybugDB. Tables: Reso…
- `env` (string): The environment whose live state the view shows. Without it, the declared source.
- `inputs` (string): With template: its inputs as a JSON object of strings ({"kind": "aws_internet_gateway"}).
- `params` (string): With cypher: its parameters, as a JSON object ({"k": "aws_s3_bucket"} for $k). Values are strings, numbers, booleans or null.
- `plan` (boolean): The plan read (behold#922): the query is answered at the plan depth (CHANT_TERRAFORM_OBSERVE=plan, chant search --live --env), and the cards carry each Terraform root's planAction, planActions, planR…
- `query` (string): A chant search query whose matches the view lights, such as `attr:internetFacing=true`.
- `resource` (string): Start from a resource (behold#1022): a name, an IR id, a block address or a cloud id (an ARN, an instance id, a Kubernetes UID). The answer is one node or the candidates, with the investigation templ…
- `show` (string): Comma-separated attributes to carry on each row (chant search --show).
- `snapshot` (string): With env: a recorded snapshot to show instead of a live read (latest, or a ref chant lists). Not with plan.
- `template` (string): With resource that resolves to one node: run one of the templates resolve lists for that node, by name. They are chant's and the project's own (.chant/templates.json), each with its origin; ask with…

### `show` (~574 tokens)

Move the person's open behold tab to a view: the query's matches lit and the rest dimmed, in the env or snapshot asked for. The tab says an agent moved it and keeps the view it left one Back away. It changes nothing but what the tab shows. Needs a running behold serve of this project; with none, it refuses and names the command that starts one. With no tab open it answers the link to hand the person. The query is chant's search grammar, unchanged (https://intentius.io/chant/cli/search/). env reads that environment live (chant search --live --env <env>); snapshot with env replays a recorded snapshot instead (chant --snapshot <ref>, such as latest); chant's --explain is always on, for the near-misses drawn. It takes no --check-live or --check-snapshot (drift is not drawn), and no --ambient, --src or --fail-on-drift. Read the capabilities tool first: it lists the capabilities a cap: term asks (cap:internet-reachable(22), ->[cap=network]), each one's meaning, parameters and rule (INTENTIUS/chant#4375).

Input parameters:

- `countBy` (string): Group the matches the way chant search --count-by does: kind, module or attr:<name> (behold#922). The answer is chant's groups and total; the view lights each group's cards and labels the group with…
- `env` (string): The environment whose live state the view shows. Without it, the declared source.
- `focus` (string): Comma-separated box ids to keep at full weight; the rest dims.
- `member` (string): Show one estate member.
- `node` (string): A card to select, by behold's node id.
- `plan` (boolean): The plan read (behold#922): the query is answered at the plan depth (CHANT_TERRAFORM_OBSERVE=plan, chant search --live --env), and the cards carry each Terraform root's planAction, planActions, planR…
- `query` (string): A chant search query whose matches the view lights, such as `attr:internetFacing=true`.
- `snapshot` (string): With env: a recorded snapshot to show instead of a live read (latest, or a ref chant lists). Not with plan.
- `zoom` (string): The zoom: components, resources, attributes, or another behold offers.

### `shot` (~609 tokens)

A PNG of a view, framed on the query's matches, and the link to the same view. Returns the image inline. Uses the running behold serve of this project; with none, it starts behold in this process for the shot and stops it after. Needs the optional peer playwright-core and a browser; without them it refuses with the install line. Writes no file. Without a query, node or focus frames the view's own selection. The query is chant's search grammar, unchanged (https://intentius.io/chant/cli/search/). env reads that environment live (chant search --live --env <env>); snapshot with env replays a recorded snapshot instead (chant --snapshot <ref>, such as latest); chant's --explain is always on, for the near-misses drawn. It takes no --check-live or --check-snapshot (drift is not drawn), and no --ambient, --src or --fail-on-drift. Read the capabilities tool first: it lists the capabilities a cap: term asks (cap:internet-reachable(22), ->[cap=network]), each one's meaning, parameters and rule (INTENTIUS/chant#4375).

Input parameters:

- `countBy` (string): Group the matches the way chant search --count-by does: kind, module or attr:<name> (behold#922). The answer is chant's groups and total; the view lights each group's cards and labels the group with…
- `env` (string): The environment whose live state the view shows. Without it, the declared source.
- `focus` (string): Comma-separated box ids to keep at full weight; the rest dims.
- `height` (integer): The page's height in px (default 1000).
- `member` (string): Show one estate member.
- `node` (string): A card to select, by behold's node id.
- `plan` (boolean): The plan read (behold#922): the query is answered at the plan depth (CHANT_TERRAFORM_OBSERVE=plan, chant search --live --env), and the cards carry each Terraform root's planAction, planActions, planR…
- `query` (string): A chant search query whose matches the view lights, such as `attr:internetFacing=true`.
- `snapshot` (string): With env: a recorded snapshot to show instead of a live read (latest, or a ref chant lists). Not with plan.
- `width` (integer): The page's width in px (default 1600).
- `zoom` (string): The zoom: components, resources, attributes, or another behold offers.

### `capabilities` (~177 tokens)

Read this first. The capability catalog chant answers cap: terms with, for the estate behold serves: chant's read-contract capabilities document, verbatim (chant search --capabilities --json, INTENTIUS/chant#4375; chant workspace search --capabilities --json on a declared workspace). Each capability has its name, what it means, its parameters, the search term that asks it (cap:internet-reachable, ->[cap=network]), the questions it answers, what it does not evaluate, its rule, and which lexicons supply it. Then ask with that term; a cap: match's path is in its because hops, and show draws it. Uses the running behold serve of this project; with none, behold answers in this process. Takes no arguments. Needs chant 0.124.46 or newer.

### `graph` (~404 tokens)

Draw the repository as a graph, or a cut of it. A host that renders MCP Apps shows the interactive view inline; every other host gets this answer: how many roots, cards and edges, the cut asked for, a Mermaid diagram when it fits (else a table of roots with card counts), and a link to the running behold serve at that view. The answer holds no document JSON, and no card list when the diagram does not fit. For a big estate ask again with around, a root or card id from the table, to see one part. Reads the files and git objects in this process; it needs no serve and no chant search. dir is a folder inside the project (default: the project). around takes comma-separated card or box ids and cuts to them and what is within depth hops (default 1); direction up follows what they depend on, down what depends on them. changed takes <base>..<head> (or <base>...<head> from the merge base) and cuts to what changed between the two commits, each card marked added, removed, changed or neighbour. around and changed are two cuts: give one.

Input parameters:

- `around` (string): Comma-separated card or box ids to cut around (a root such as a Terraform root, or a card id from the roots table), with what is within depth hops of them.
- `changed` (string): A revision range, <base>..<head> or <base>...<head> (from the merge base): cut to what changed between the two commits and the cards within depth hops.
- `depth` (integer): Hops from the cards around names, or from what changed (default 1).
- `dir` (string): A folder inside the project to read, relative to it (default: the project folder).
- `direction` (string): With around: up follows what the cards depend on, down what depends on them (default both).

### `graph_document` (~220 tokens)

For the graph view only, not for a model: the behold-graph/1 document of a graph call, taking the same arguments, so the view can draw it. The model is never offered this tool, and its answer never enters the conversation.

Input parameters:

- `around` (string): Comma-separated card or box ids to cut around (a root such as a Terraform root, or a card id from the roots table), with what is within depth hops of them.
- `changed` (string): A revision range, <base>..<head> or <base>...<head> (from the merge base): cut to what changed between the two commits and the cards within depth hops.
- `depth` (integer): Hops from the cards around names, or from what changed (default 1).
- `dir` (string): A folder inside the project to read, relative to it (default: the project folder).
- `direction` (string): With around: up follows what the cards depend on, down what depends on them (default both).

## Diagnostics

Captured diagnostic sections: Provenance, Vulnerabilities, Dependencies. The full working is on the page: https://verifymcp.io/servers/intentius-behold/intentius-behold#diagnostics

## Score history

- 2026-11-22: 77
- 2026-11-21: 62
- 2026-11-20: 76
- 2026-11-19: 76
- 2026-11-18: 60
- 2026-11-17: 75
- 2026-11-16: 54
- 2026-11-15: 59

## Common questions

### What is the behold MCP server?

behold is an MCP server listed in the public MCP registry as io.github.INTENTIUS/behold. Your infrastructure repo as a graph, from the files. Draws Terraform roots, no state needed. This page covers its npm package (@intentius/behold).

### Is the behold MCP server safe to use?

behold scores 77 out of 100 on VerifyMCP. We recorded 5 known advisories against it as of 22 November 2026. It declares no install or post-install scripts. Its build provenance is signed and verified. That is a record of what we were able to check automatically, not an endorsement. The category breakdown on this page shows every signal behind the number, including the ones we could not confirm.

### What tools does the behold MCP server expose?

behold exposes 6 tools: ask, show, shot, capabilities, graph, graph_document. Their descriptions and schemas cost roughly 3,197 tokens of context every time the server is loaded.

### Is the behold MCP server still maintained?

behold is still listed as active in the MCP registry. We last reached this channel on 22 November 2026. Those dates come from our own scans of the registry and the channel itself, not from anything the publisher announced.

### What licence is the behold MCP server under?

behold declares the Apache-2.0 licence, which is OSI-approved. That covers the source only, and says nothing about the cost of any service it calls.

### What did VerifyMCP test on the behold MCP server?

VerifyMCP ran 17 checks on behold, covering Malware scan, Known CVEs, Install scripts, Dependency health and 13 more. Endpoint Security and Transport & Reachability do not apply to an npm package, so they were not tested. The score does not measure everything, and the scoring documentation lists what it leaves out.

## Links

- npm package: https://www.npmjs.com/package/@intentius/behold
- Socket report: https://socket.dev/npm/package/@intentius/behold
- Repository: https://github.com/INTENTIUS/behold
- Changelog RSS feed: https://verifymcp.io/servers/intentius-behold/intentius-behold.xml
- Changelog JSON feed: https://verifymcp.io/servers/intentius-behold/intentius-behold/changelog.json
- Trust report (JSON): https://verifymcp.io/servers/intentius-behold/intentius-behold.json
- HTML version of this page: https://verifymcp.io/servers/intentius-behold/intentius-behold
