# io.github.Shoberman2/clipswarm (npm · clipswarm)

Clip YouTube videos and livestreams into viral 9:16 shorts, in parallel, for AI agents

- Trust score: 60/100 (medium)
- Registry status: active
- Liveness: live
- Owner verified: no
- Last scored: 2026-10-17

## Components

- npm · `clipswarm`: 60/100 (this document), [markdown](https://verifymcp.io/servers/shoberman2-clipswarm/clipswarm.md), [page](https://verifymcp.io/servers/shoberman2-clipswarm/clipswarm)

## Channel facts

- Registry: `npm`
- Package: `clipswarm`
- Version: `0.1.0`
- 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 2026-10-17.

- **Supply Chain Security**: 98/100
  - No malware found by supply-chain analysis.
  - No known CVEs affecting this package version or its production dependencies.
  - No install/post-install scripts declared.
  - 32 of 108 dependencies flagged as unhealthy.
- **Provenance & Transparency**: 19/100
  - Repository check failed: the declared repository URL redirects; it must resolve directly.
  - Provenance check failed: no build-provenance attestation is published.
  - Clear OSI-approved license (MIT).
  - Actively maintained (last published 7 days ago).
  - Disclosure check failed: no security disclosure policy was found in the source repository.
- **Schema Quality & AI Usability**: 66/100
  - AI-judged instruction clarity (good).
  - Tool/resource definitions use about 1267 tokens (~105/item across 12 items; 12 tools + 0 resources), lean.
  - Usage-examples check failed: none of the tools include examples.
- **Stability & Change Management**: 0/100
  - Stability not yet verified: not enough scan history yet (needs a 30-day window).
- **Tool Coverage**: 86/100
  - 100% of tools have a non-trivial description (not blank, and not just the tool's name).
  - 59% of tool parameters carry a description.
- **Tool Safety**: 75/100
  - No prompt-injection markers were found in the server instructions, tool names or descriptions we captured.
  - 0 of 1 tool(s) whose name or description implies an irreversible operation declare an MCP destructiveHint annotation; "delete_watch" implies "delete" and declares no destructiveHint at all, which the MCP spec reads as destructive by default.
  - An AI judge read all 12 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.

**Unverified: 1 category.** A category scored 0 because we could not verify it: a data source with nothing on this package, evidence we could not reach, or a check we could not run. We only credit what we can confirm.

## Install

### How do I install the io.github.Shoberman2/clipswarm MCP server?

io.github.Shoberman2/clipswarm runs locally as an npm package, launched with npx -y clipswarm. 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 shoberman2-clipswarm -- npx -y clipswarm
```

### Cursor

```json
{
  "mcpServers": {
    "shoberman2-clipswarm": {
      "command": "npx",
      "args": [
        "-y",
        "clipswarm"
      ]
    }
  }
}
```

### VS Code

```json
{
  "servers": {
    "shoberman2-clipswarm": {
      "command": "npx",
      "args": [
        "-y",
        "clipswarm"
      ]
    }
  }
}
```

### Codex

```bash
codex mcp add shoberman2-clipswarm -- npx -y clipswarm
```

### opencode

```json
{
  "$schema": "https://opencode.ai/config.json",
  "mcp": {
    "shoberman2-clipswarm": {
      "type": "local",
      "command": [
        "npx",
        "-y",
        "clipswarm"
      ],
      "enabled": true
    }
  }
}
```

### OpenClaw

```bash
openclaw mcp add shoberman2-clipswarm --command npx --arg -y --arg clipswarm
```

### Hermes

```yaml
mcp_servers:
  shoberman2-clipswarm:
    command: "npx"
    args: ["-y", "clipswarm"]
```

### Netclaw

```json
{
  "McpServers": {
    "shoberman2-clipswarm": {
      "Transport": "stdio",
      "Command": "npx",
      "Arguments": [
        "-y",
        "clipswarm"
      ]
    }
  }
}
```

### Vellum

```bash
assistant mcp add shoberman2-clipswarm -t stdio -c npx -a -y clipswarm
```

### Other

```json
{
  "mcpServers": {
    "shoberman2-clipswarm": {
      "command": "npx",
      "args": [
        "-y",
        "clipswarm"
      ]
    }
  }
}
```

## 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.

### 2026-10-14 (score 60, +27)

- [security improvement] Injection markers: unverified → pass
- [security] First check of Judged manipulation: pass
- [security] Stability: Stability not yet verified: not enough scan history yet (needs a 30-day window).
- [functional improvement] Tool coverage: unverified → 100
- [functional improvement] MCP protocol: unverified → pass
- [functional] First check of Schema quality: fail
- [functional] First check of Tool coverage: 59
- [functional] First check of Destructive annotations: 0
- [functional] First check of Schema quality: pass
- [functional] First check of Schema quality: good

### 2026-10-12 (score 33, −5)

- [security regression] Source repository: pass → fail

### 2026-10-11 (score 38)

First indexed and scored.

## MCP tools (12)

### `get_video_info` (~60 tokens)

Get a YouTube video's title, channel, duration, chapters and liveStatus ("is_live" means it's streaming now: clip it with times relative to "now"). Use chapters to find natural clip boundaries.

Input parameters:

- `url` (string, required): YouTube URL

### `get_transcript` (~222 tokens)

Get the timestamped transcript of a YouTube video. Optionally limit to a time window to keep output small. Read it to decide which moments to clip.

Input parameters:

- `from` (string|number): Seconds (83) or "1:23" / "1:02:03.5". Negative = from the end ("-0:30"). For LIVE streams: "now", negative offsets from the live edge (-90), or ISO time ("2026-10-08T00:40:00Z"); roughly the last hou…
- `lang` (string): Caption language code, default "en"
- `to` (string|number): Seconds (83) or "1:23" / "1:02:03.5". Negative = from the end ("-0:30"). For LIVE streams: "now", negative offsets from the live edge (-90), or ISO time ("2026-10-08T00:40:00Z"); roughly the last hou…
- `url` (string, required)

### `search_transcript` (~93 tokens)

Find where a phrase is said in a YouTube video. Returns time windows with surrounding text, ready to pass to create_clips.

Input parameters:

- `contextSec` (number): Seconds of padding around each hit, default 15
- `lang` (string)
- `limit` (number)
- `query` (string, required): Words to find (all must appear in the caption line)
- `url` (string, required)

### `create_clips` (~166 tokens)

Cut one or more clips from YouTube videos (including streams that are live right now) in parallel. Only the requested sections are downloaded. Jobs can span different videos. Returns a result per job; failures don't abort the batch. Live results include the wall-clock range covered. Viral clips take ~15-60s each; if the batch isn't done within waitSec, you get finished results plus a jobId: call get_clips with it until pending is 0.

Input parameters:

- `clips` (array, required)
- `outDir` (string): Output directory, default /workdir/clips
- `waitSec` (number): Seconds to wait before returning, default 40. If clips are still rendering, the response includes a jobId; collect the rest with get_clips.

### `get_clips` (~84 tokens)

Get results of a create_clips batch that was still running (create_clips returns a jobId when it doesn't finish within waitSec). Waits up to waitSec for more clips to finish. Call repeatedly until pending is 0.

Input parameters:

- `jobId` (string, required)
- `waitSec` (number): Seconds to wait for completion before returning, default 40

### `create_watch` (~341 tokens)

Watch a category: on a schedule, search YouTube for new videos about it and auto-clip the best moments into viral shorts (moments and hooks picked by the local Claude Code CLI if installed, else YouTube's most-replayed peaks, else a transcript heuristic). Anything it can't clip (and live streams, if enabled) goes to the inbox for you. Replaces a watch with the same id. Runs via the background scheduler (see watch_scheduler).

Input parameters:

- `autoClip` (boolean): Default true. false = only collect videos into the inbox
- `clipSeconds` (number): Target clip length, default 45
- `clipsPerVideo` (number): Default 2
- `every` (string, required): How often to check: "30m", "6h", "1d" (min 5m)
- `id` (string): Short name; defaults to a slug of the query
- `includeLive` (boolean): Also surface streams that are live right now
- `maxAgeHours` (number): Only videos uploaded within this many hours (default 72)
- `maxDurationSec` (number): Default 14400
- `maxVideosPerRun` (number): Default 3
- `minDurationSec` (number): Default 120 (skips Shorts)
- `minViews` (number)
- `outDir` (string): Default ~/clipswarm/<id>
- `picker` (string): How moments are chosen. Default "auto"
- `query` (string, required): The category, phrased like a YouTube search, e.g. "netflix stock analysis"
- `style` (string)

### `list_watches` (~24 tokens)

List watches with their schedule, last and next run, plus scheduler status.

### `delete_watch` (~21 tokens)

Stop and remove a watch.

Input parameters:

- `id` (string, required)

### `run_watch` (~45 tokens)

Run a watch right now instead of waiting for its schedule. Runs in the background (searching and rendering takes minutes); check get_inbox afterwards.

Input parameters:

- `id` (string, required)

### `get_inbox` (~105 tokens)

Videos found by watches. "clipped" = auto-clipped (paths in clips; note says how moments were picked). "needs_agent" = couldn't be clipped automatically: read the transcript, pick moments, call create_clips, then mark it "done". "live" = streaming now. "failed" = see note.

Input parameters:

- `limit` (number): Most recent N, default 30
- `status` (string)
- `watchId` (string)

### `update_inbox_item` (~49 tokens)

Mark an inbox video as handled ("done") or not worth clipping ("dismissed").

Input parameters:

- `note` (string)
- `status` (string, required)
- `videoId` (string, required)

### `watch_scheduler` (~57 tokens)

Status, install or uninstall the background job that runs due watches every 5 minutes (launchd on macOS; returns a crontab line elsewhere). Install only when the user wants watches to run automatically.

Input parameters:

- `action` (string, required)

## Diagnostics

Captured diagnostic sections: Provenance, Dependencies. The full working is on the page: https://verifymcp.io/servers/shoberman2-clipswarm/clipswarm#diagnostics

## Score history

- 2026-10-17: 60
- 2026-10-16: 60
- 2026-10-15: 60
- 2026-10-14: 60
- 2026-10-13: 33
- 2026-10-12: 33
- 2026-10-11: 38

## Common questions

### What is the io.github.Shoberman2/clipswarm MCP server?

io.github.Shoberman2/clipswarm is an MCP server listed in the public MCP registry as io.github.Shoberman2/clipswarm. Clip YouTube videos and livestreams into viral 9:16 shorts, in parallel, for AI agents. This page covers its npm package (clipswarm).

### Is the io.github.Shoberman2/clipswarm MCP server safe to use?

io.github.Shoberman2/clipswarm scores 60 out of 100 on VerifyMCP. We found no known CVEs affecting it as of 17 October 2026. It declares no install or post-install scripts. 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 io.github.Shoberman2/clipswarm MCP server expose?

io.github.Shoberman2/clipswarm exposes 12 tools: get_video_info, get_transcript, search_transcript, create_clips, get_clips, and 7 more. Their descriptions and schemas cost roughly 1,267 tokens of context every time the server is loaded.

### Is the io.github.Shoberman2/clipswarm MCP server still maintained?

io.github.Shoberman2/clipswarm is still listed as active in the MCP registry. We last reached this channel on 17 October 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 io.github.Shoberman2/clipswarm MCP server under?

io.github.Shoberman2/clipswarm declares the MIT licence, which is OSI-approved. That covers the source only, and says nothing about the cost of any service it calls.

## Links

- npm package: https://www.npmjs.com/package/clipswarm
- Socket report: https://socket.dev/npm/package/clipswarm
- Changelog RSS feed: https://verifymcp.io/servers/shoberman2-clipswarm/clipswarm.xml
- Changelog JSON feed: https://verifymcp.io/servers/shoberman2-clipswarm/clipswarm.json
- HTML version of this page: https://verifymcp.io/servers/shoberman2-clipswarm/clipswarm