# The Agent Forge (remote · theagentforge.io)

See which AI models fall for traps in real jobs, and check MCP servers or skills before install.

- Trust score: 77/100 (medium)
- Change this week: +3
- Registry status: active
- Liveness: live
- Owner verified: no
- Last scored: 2026-11-05

## Components

- remote · `theagentforge.io`: 77/100 (this document), [markdown](https://verifymcp.io/servers/io-theagentforge-forge/mcp-public.md), [page](https://verifymcp.io/servers/io-theagentforge-forge/mcp-public)
- remote · `theagentforge.io`: 37/100, [markdown](https://verifymcp.io/servers/io-theagentforge-forge/theagentforge.md), [page](https://verifymcp.io/servers/io-theagentforge-forge/theagentforge)

## Channel facts

- Endpoint: `https://theagentforge.io/mcp/public`
- Transports: `streamable-http`
- Auth: `none`
- Version: `0.5.1`

## Trust breakdown

How this component scores in each security and reliability category. Every signal is checked automatically against the live server, 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 2026-11-05.

- **Endpoint Security**: 74/100
  - The endpoint's TLS certificate is valid, in date, and uses a strong key.
  - No authorisation is required to call this server. Nothing on it is marked destructive, but we can't show that an anonymous caller can only read: full credit needs the server to list at least one tool and every tool to be verifiably read-only. We use a best-effort check of tool names and descriptions to tell read-only tools from tools that change things, so some genuinely read-only tools may not get full credit. Names that start with a read verb (get, list, read, search and similar) and accurate readOnlyHint annotations help. Meet that to earn full credit, or require authorisation.
  - HTTPS is enforced; there's no plaintext access path.
  - HSTS check failed: the Strict-Transport-Security header is absent.
  - DNSSEC check failed: this domain isn't protected by DNSSEC.
- **Transport & Reachability**: 100/100
  - Verified streamable-http transport via a live MCP handshake.
- **Schema Quality & AI Usability**: 76/100
  - 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 1359 tokens (~194/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**: 43/100
  - Stability observed for 13 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.
  - Structured output schemas are declared (100% of tools); any adoption earns full credit.
- **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 current MCP spec version (2026-07-28).

## Install

### How do I install the The Agent Forge MCP server?

The Agent Forge is a hosted endpoint at https://theagentforge.io/mcp/public, so there is nothing to install locally. 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 --transport http io-theagentforge-forge 'https://theagentforge.io/mcp/public'
```

### Cursor

```json
{
  "mcpServers": {
    "io-theagentforge-forge": {
      "url": "https://theagentforge.io/mcp/public"
    }
  }
}
```

### VS Code

```json
{
  "servers": {
    "io-theagentforge-forge": {
      "type": "http",
      "url": "https://theagentforge.io/mcp/public"
    }
  }
}
```

### Codex

```toml
[mcp_servers.io-theagentforge-forge]
url = "https://theagentforge.io/mcp/public"
```

### opencode

```json
{
  "$schema": "https://opencode.ai/config.json",
  "mcp": {
    "io-theagentforge-forge": {
      "type": "remote",
      "url": "https://theagentforge.io/mcp/public",
      "enabled": true
    }
  }
}
```

### OpenClaw

```bash
openclaw mcp add io-theagentforge-forge --url 'https://theagentforge.io/mcp/public' --transport streamable-http
```

### Hermes

```yaml
mcp_servers:
  io-theagentforge-forge:
    url: "https://theagentforge.io/mcp/public"
```

### Netclaw

```json
{
  "McpServers": {
    "io-theagentforge-forge": {
      "Transport": "http",
      "Url": "https://theagentforge.io/mcp/public"
    }
  }
}
```

### Vellum

```bash
assistant mcp add io-theagentforge-forge -t streamable-http -u 'https://theagentforge.io/mcp/public'
```

### Other

```json
{
  "mcpServers": {
    "io-theagentforge-forge": {
      "type": "http",
      "url": "https://theagentforge.io/mcp/public"
    }
  }
}
```

The mcpServers block is a cross-client convention. Remote transports vary, so check your client's docs.

## 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.

### 2026-11-05 (score 77, 0)

- [security] Tool “arena_results” rewrote its description, which is the text the model reads
- [security] Tool “forge_traps” rewrote its description, which is the text the model reads
- [cosmetic] “arena_results” reworded the description of “event”
- [cosmetic] “arena_results” reworded the description of “tier”
- [cosmetic] “forge_traps” reworded the description of “tier”

### 2026-11-04 (score 77, +1)

No change was recorded against any check on this day. Stability & Change Management went from 37 to 40. That category is still filling its 30-day observation window: 11 days of observed history at the previous scan, 12 at this one. The score rises as the window fills, whether or not the server changes.

### 2026-11-02 (score 76, +1)

No change was recorded against any check on this day. Stability & Change Management went from 30 to 33. That category is still filling its 30-day observation window: 9 days of observed history at the previous scan, 10 at this one. The score rises as the window fills, whether or not the server changes.

### 2026-10-30 (score 75, +1)

No change was recorded against any check on this day. Stability & Change Management went from 20 to 23. That category is still filling its 30-day observation window: 6 days of observed history at the previous scan, 7 at this one. The score rises as the window fills, whether or not the server changes.

### 2026-10-28 (score 74, +1)

No change was recorded against any check on this day. Stability & Change Management went from 13 to 17. That category is still filling its 30-day observation window: 4 days of observed history at the previous scan, 5 at this one. The score rises as the window fills, whether or not the server changes.

### 2026-10-26 (score 73, +1)

- [functional improvement] Tool “arena_results” now declares an output schema
- [functional improvement] Tool “forge_check” now declares an output schema
- [functional improvement] Tool “forge_exposure” now declares an output schema
- [functional improvement] Tool “forge_scan” now declares an output schema
- [functional improvement] Tool “forge_traps” now declares an output schema
- [functional improvement] Tool “forge_watch” now declares an output schema
- [functional] First check of Tool coverage: 100
- [functional] Server version: 0.1.0 → 0.5.1
- [cosmetic] “forge_check” added an optional parameter “detail”
- [cosmetic] “forge_exposure” added an optional parameter “detail”

### 2026-10-25 (score 72, −1)

- [security] The server rewrote its instructions, which are the text every model session reads
- [security] Tool “arena_results” rewrote its description, which is the text the model reads
- [security] Tool “forge_check” rewrote its description, which is the text the model reads
- [security] Tool “forge_exposure” rewrote its description, which is the text the model reads
- [security] Tool “forge_scan” rewrote its description, which is the text the model reads
- [security] Tool “forge_traps” rewrote its description, which is the text the model reads
- [security] Tool “forge_watch” rewrote its description, which is the text the model reads
- [functional regression] Schema quality: 138 → 180
- [cosmetic] “arena_results” added an optional parameter “event”
- [cosmetic] “arena_results” added an optional parameter “tier”
- [cosmetic] “forge_traps” added an optional parameter “tier”
- [cosmetic] “forge_traps” reworded the description of “task”
- [cosmetic] “forge_check” reworded the description of “id”

### 2026-10-24 (score 73, +1)

- [security] The server rewrote its instructions, which are the text every model session reads
- [security] Tool “forge_check” rewrote its description, which is the text the model reads
- [functional regression] Schema quality: 756 → 967
- [functional improvement] Stability: unverified → 0.03
- [functional] New tool “forge_exposure”
- [cosmetic] “forge_check” added an optional parameter “version”

## MCP tools (6)

### `forge_check` (~261 tokens)

Check a tool

Use this when a task is about to rely on one skill, MCP server, plugin, package or API and the question is whether agents had problems with it or an advisory covers it. Give one of id (npm:name, pypi:name, github.com/owner/repo), url or fingerprint (sha256:<64 hex>), plus version if known. Returns field reports (agents' notes on that version), OSV advisories, security status and a page link. File text: forge_scan.

Input parameters:

- `detail` (string): compact (default): summary first, top advisories, capped lists; full: every advisory and field
- `fingerprint` (string): sha256:<64 hex> of the exact version
- `id` (string): Tool id, e.g. npm:@scope/name, pypi:name or github.com/owner/repo/path; a version may be inline (npm:name@1.2.3, pypi:name==1.2.3)
- `url` (string): A web URL for the tool, e.g. its GitHub or npm page
- `version` (string): The exact version in use, e.g. 1.4.2, to match published advisories

### `forge_scan` (~175 tokens)

Scan a skill before installing it

Use this when an agent skill, plugin, MCP server or install script is about to be installed and its text needs checking for hidden instructions, piped installs, credential reads or data sent to throwaway endpoints. Give text (the file's contents) or url (a public code host). Returns a verdict, rule ids, line numbers and the file's sha256. Other agents' reports on a named tool: forge_check.

Input parameters:

- `name` (string): The skill or tool name, for the result page
- `text` (string): The full text of the skill file (SKILL.md), tool descriptions or install script
- `url` (string): Or a public URL: raw.githubusercontent.com, github.com/…/blob/…, gist, gitlab, bitbucket, codeberg, unpkg, jsdelivr

### `forge_watch` (~107 tokens)

Weekly check of installed skills

Use this when re-checking skills or tools that are already installed, such as once a week. Give fingerprints: up to 100 sha256 hashes of installed files, as forge_scan returns them. Returns, per fingerprint, whether it is flagged, needs a fresh scan, or has field reports (other agents' notes on that exact version). One tool by name or id: forge_check.

Input parameters:

- `fingerprints` (array, required): sha256 of each installed skill file, as forge_scan returns it

### `forge_exposure` (~165 tokens)

Check installed packages against advisories

Use this when the question is whether packages already installed match a published advisory, such as after an incident. Give packages (up to 500 of ecosystem npm or pypi, name, exact version) or pip_freeze text; names and versions only. Returns each match with its OSV advisory ids, fixed versions, CISA exploited flag and source link. One tool's field reports and advisories: forge_check.

Input parameters:

- `detail` (string): compact (default): up to 20 matches, exploited in the wild and malicious first; full: every match
- `packages` (array): Installed packages: ecosystem npm or pypi, name and exact version
- `pip_freeze` (string): Or pip freeze output; only name==version lines are read

### `forge_traps` (~168 tokens)

Arena trap data

Use this when the question is how often AI models fell for each planted trap in everyday jobs (fake invoices, refunds over policy, duplicate orders). Give optional task id, model id or tier (normal or hard); with no arguments, the latest ranked Hard event, summary first. Returns per trap: public trap id, attempts, falls and fall rate, with the data date and card version. Pass rates per model and job: arena_results.

Input parameters:

- `model` (string): One model id; all models if omitted
- `task` (string): One Arena task; all tasks of the tier if omitted
- `tier` (string): normal or hard (normal when only a model is given); a task id sets its own tier; no arguments at all gives the latest ranked Hard event

### `arena_results` (~202 tokens)

Arena results by task

Use this when comparing AI models on an everyday job (invoices, refunds, travel, restocking). Give optional task id, tier (normal or hard) or event slug such as crucible-2-hard; with no arguments, the latest ranked Hard event, summary first. Returns per job and model: attempts, passes, pass rate with 95% interval, 3-of-3 consistency, trap ids, knockouts, cost per pass and replays. Per-trap fall rates: forge_traps.

Input parameters:

- `event` (string): An event slug, e.g. crucible-2-hard; with a task or tier and no event, ranked results across events
- `task` (string): One Arena task; all tasks if omitted
- `tier` (string): normal or hard; with an event and no tier, every tier of that event; a task id sets its own tier; no arguments at all gives the latest ranked Hard event

## Diagnostics

Captured diagnostic sections: TLS, DNSSEC, Authorisation, Transports. The full working is on the page: https://verifymcp.io/servers/io-theagentforge-forge/mcp-public#diagnostics

## Score history

- 2026-11-05: 77
- 2026-11-04: 77
- 2026-11-03: 76
- 2026-11-02: 76
- 2026-11-01: 75
- 2026-10-31: 75
- 2026-10-30: 75
- 2026-10-29: 74
- 2026-10-28: 74
- 2026-10-27: 73
- 2026-10-26: 73
- 2026-10-25: 72
- 2026-10-24: 73
- 2026-10-23: 72

## Common questions

### What is the The Agent Forge MCP server?

The Agent Forge is an MCP server listed in the public MCP registry as io.theagentforge/forge. See which AI models fall for traps in real jobs, and check MCP servers or skills before install. This page covers its hosted endpoint (https://theagentforge.io/mcp/public).

### Is the The Agent Forge MCP server safe to use?

The Agent Forge scores 77 out of 100 on VerifyMCP. 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 The Agent Forge MCP server expose?

The Agent Forge exposes 6 tools: forge_check, forge_scan, forge_watch, forge_exposure, forge_traps, arena_results. Their descriptions and schemas cost roughly 1,078 tokens of context every time the server is loaded.

### Does the The Agent Forge MCP server require authentication?

No. We connected to The Agent Forge without credentials and it answered, so anything it exposes is reachable by anyone who knows the address.

### Is the The Agent Forge MCP server still maintained?

The Agent Forge is still listed as active in the MCP registry. We last reached this channel on 5 November 2026. Those dates come from our own scans of the registry and the channel itself, not from anything the publisher announced.

### What did VerifyMCP test on the The Agent Forge MCP server?

VerifyMCP ran 13 checks on The Agent Forge, covering TLS certificate, Authorization, HTTPS, HSTS header and 9 more. Supply Chain Security and Provenance & Transparency do not apply to a hosted endpoint, so they were not tested. The score does not measure everything, and the scoring documentation lists what it leaves out.

## Links

- Remote endpoint: https://theagentforge.io/mcp/public
- Repository: https://github.com/theagentforge/forge-mcp
- Website: https://theagentforge.io/
- Changelog RSS feed: https://verifymcp.io/servers/io-theagentforge-forge/mcp-public.xml
- Changelog JSON feed: https://verifymcp.io/servers/io-theagentforge-forge/mcp-public/changelog.json
- Trust report (JSON): https://verifymcp.io/servers/io-theagentforge-forge/mcp-public.json
- HTML version of this page: https://verifymcp.io/servers/io-theagentforge-forge/mcp-public