# pepe (oci · ghcr.io/omarmhaimdat/pepe)

Load-test a URL, ping it by phase, ramp to find the capacity, compare runs: JSON reports, guardrails

- Trust score: 46/100 (low)
- Change this week: +3
- Registry status: active
- Liveness: live
- Owner verified: no
- Last scored: 2026-11-05

## Components

- oci · `ghcr.io/omarmhaimdat/pepe`: 46/100 (this document), [markdown](https://verifymcp.io/servers/omarmhaimdat-pepe/ghcr-io-omarmhaimdat-pepe.md), [page](https://verifymcp.io/servers/omarmhaimdat-pepe/ghcr-io-omarmhaimdat-pepe)
- pypi · `pepe-http`: 35/100, [markdown](https://verifymcp.io/servers/omarmhaimdat-pepe/pepe-http.md), [page](https://verifymcp.io/servers/omarmhaimdat-pepe/pepe-http)

## Channel facts

- Registry: `oci`
- Package: `ghcr.io/omarmhaimdat/pepe`
- Version: `0.57.0`
- 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 2026-11-05.

- **Supply Chain Security**: 0/100
  - No malware scan is available for this kind of package: the supply-chain vendors we use do not cover it. This is a permanent gap in our coverage, not a finding about the package.
  - Known CVEs could not be checked: this artifact ships no SBOM or dependency manifest, so there is no dependency list to read.
  - Install-script risk not yet assessed.
  - Dependency health could not be checked: this artifact ships no SBOM or dependency manifest, so there is no dependency list to read.
- **Provenance & Transparency**: 48/100
  - 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 9 days ago).
  - Publishes a security disclosure policy (SECURITY.md).
- **Schema Quality & AI Usability**: 69/100
  - 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 3787 tokens (~473/item across 8 items; 8 tools + 0 resources), over budget; trim descriptions and params.
  - Usage-examples check failed: none of the tools include examples.
- **Stability & Change Management**: 73/100
  - Stability observed for 22 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**: 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; "replay" implies "send" and declares no destructiveHint at all, which the MCP spec reads as destructive by default.
  - An AI judge read all 9 captured unit(s) of tool text and found none that tries to manipulate the model reading it.
- **Capabilities**: 40/100
  - Spec-recency check failed: implements MCP spec 2025-03-26; 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 pepe MCP server?

pepe runs locally as a container image, launched with docker run --rm -i ghcr.io/omarmhaimdat/pepe:0.57.0. Ready-made configuration for Claude, Cursor, VS Code, Codex and 3 more is on this page, copied from each client's own documentation.

### Claude

```bash
claude mcp add omarmhaimdat-pepe -- docker run --rm -i ghcr.io/omarmhaimdat/pepe:0.57.0
```

### Cursor

```json
{
  "mcpServers": {
    "omarmhaimdat-pepe": {
      "command": "docker",
      "args": [
        "run",
        "--rm",
        "-i",
        "ghcr.io/omarmhaimdat/pepe:0.57.0"
      ]
    }
  }
}
```

### VS Code

```json
{
  "servers": {
    "omarmhaimdat-pepe": {
      "command": "docker",
      "args": [
        "run",
        "--rm",
        "-i",
        "ghcr.io/omarmhaimdat/pepe:0.57.0"
      ]
    }
  }
}
```

### Codex

```bash
codex mcp add omarmhaimdat-pepe -- docker run --rm -i ghcr.io/omarmhaimdat/pepe:0.57.0
```

### opencode

```json
{
  "$schema": "https://opencode.ai/config.json",
  "mcp": {
    "omarmhaimdat-pepe": {
      "type": "local",
      "command": [
        "docker",
        "run",
        "--rm",
        "-i",
        "ghcr.io/omarmhaimdat/pepe:0.57.0"
      ],
      "enabled": true
    }
  }
}
```

### Hermes

```yaml
mcp_servers:
  omarmhaimdat-pepe:
    command: "docker"
    args: ["run", "--rm", "-i", "ghcr.io/omarmhaimdat/pepe:0.57.0"]
```

### Netclaw

```json
{
  "McpServers": {
    "omarmhaimdat-pepe": {
      "Transport": "stdio",
      "Command": "docker",
      "Arguments": [
        "run",
        "--rm",
        "-i",
        "ghcr.io/omarmhaimdat/pepe:0.57.0"
      ]
    }
  }
}
```

### Other

```json
{
  "mcpServers": {
    "omarmhaimdat-pepe": {
      "command": "docker",
      "args": [
        "run",
        "--rm",
        "-i",
        "ghcr.io/omarmhaimdat/pepe:0.57.0"
      ]
    }
  }
}
```

## 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.

### 2026-11-03 (score 46, +1)

No change was recorded against any check on this day. Stability & Change Management went from 63 to 67. That category is still filling its 30-day observation window: 19 days of observed history at the previous scan, 20 at this one. The score rises as the window fills, whether or not the server changes.

### 2026-11-01 (score 45, +1)

No change was recorded against any check on this day. Stability & Change Management went from 57 to 60. That category is still filling its 30-day observation window: 17 days of observed history at the previous scan, 18 at this one. The score rises as the window fills, whether or not the server changes.

### 2026-10-30 (score 44, +1)

No change was recorded against any check on this day. Stability & Change Management went from 50 to 53. That category is still filling its 30-day observation window: 15 days of observed history at the previous scan, 16 at this one. The score rises as the window fills, whether or not the server changes.

### 2026-10-28 (score 43, +1)

No change was recorded against any check on this day. Stability & Change Management went from 43 to 47. That category is still filling its 30-day observation window: 13 days of observed history at the previous scan, 14 at this one. The score rises as the window fills, whether or not the server changes.

### 2026-10-26 (score 42, +1)

- [functional regression] Schema quality: 426 → 473
- [functional] Package version: 0.55.2 → 0.57.0
- [functional] Package version: 0.55.2 → 0.56.0

### 2026-10-24 (score 41, 0)

- [security regression] Provenance: fail → unverified
- [functional] Package version: 0.55.0 → 0.55.2
- [functional] Package version: 0.55.0 → 0.55.1

### 2026-10-23 (score 41, +1)

- [functional regression] Schema quality: 3012 → 3415
- [functional] Package version: 0.53.0 → 0.55.0
- [functional] Package version: 0.53.0 → 0.54.0

### 2026-10-22 (score 40, +1)

- [functional] Package version: 0.52.0 → 0.53.0

## MCP tools (8)

### `run` (~569 tokens)

Load-test a URL: send requests at a concurrency, for a count or a duration, and report throughput, latency percentiles and phases, status codes, failures by cause and anomalies. The report's `summary` has the numbers; `generator.peak_busy_percent` near 100 means pepe was the limit. With `otlp`, each request carries a traceparent and the report's `spans` say where the server's time went, by operation (a query, a call, the handler's own code).

Input parameters:

- `body` (string): The request body
- `concurrency` (integer): Requests in flight at once (one per core)
- `dry_run` (boolean): Say what would be sent and send nothing
- `duration` (string): Run for this long instead: '30s', '2m'
- `fail_if` (array): Conditions that make the exit code 4: 'p99 > 300ms', 'errors > 1%'
- `headers` (array): Headers, each 'Name: value'
- `insecure` (boolean): Accept invalid TLS certificates
- `method` (string): HTTP method; GET by default
- `no_compression` (boolean): Don't ask for a compressed body
- `no_keepalive` (boolean): A new connection for every request
- `no_redirects` (boolean): Take a redirect as the answer rather than follow it
- `otlp` (string): Take the service's OpenTelemetry spans here, as ':4318'; the service exports to http://localhost:4318
- `rate` (number): Start this many requests a second, spread evenly; concurrency is then the most in flight
- `requests` (integer): Requests to send (100); or duration
- `save` (string): Also write the report to this file, to compare against later
- `slack` (string|boolean): Keep a Slack message up to date as the run goes: true for the channel `pepe slack connect` kept, or '#channel'
- `timeout` (integer): Seconds to wait for a response (20)
- `url` (string, required): The URL to send to
- `user_agent` (string): The User-Agent to send
- `vars` (string): A CSV file whose first line names the {{holes}} in the URL, headers and body; each request takes the next row
- `warmup` (string): Send for this long before measuring: '5s'
- `webhook` (array): Send the verdict here when it ends: a Slack, Discord, Teams or Google Chat webhook, or any URL for the JSON report
- `webhook_on` (string): When the webhooks are sent, comma-separated: start, progress, anomaly, end, fail (end)

### `ping` (~458 tokens)

Ping URLs over HTTP: a request every interval to each target, each split into DNS, connect, TLS, first byte and download, with loss, jitter, percentiles, the TLS session, the certificate's expiry, and findings that say what to look at. Five pings by default.

Input parameters:

- `all_ips` (boolean): Ping every address a name resolves to
- `body` (string): The request body
- `count` (integer): Pings per target (5)
- `dry_run` (boolean): Say what would be sent and send nothing
- `duration` (string): Ping for this long instead: '30s'
- `every` (string): Time between pings: '1s', '200ms'
- `headers` (array): Headers, each 'Name: value'
- `insecure` (boolean): Accept invalid TLS certificates
- `keep_alive` (boolean): Keep the connection between pings
- `method` (string): HTTP method; GET by default
- `no_compression` (boolean): Don't ask for a compressed body
- `no_keepalive` (boolean): A new connection for every request
- `no_redirects` (boolean): Take a redirect as the answer rather than follow it
- `save` (string): Also write the report to this file, to compare against later
- `slack` (string|boolean): Keep a Slack message up to date as the run goes: true for the channel `pepe slack connect` kept, or '#channel'
- `slo` (string): Limits in milliseconds, exit code 4 when broken: 'total=500,ttfb=200'
- `targets` (array, required): URLs or hosts; several are reported apart
- `tcp` (boolean): Only connect to the port; no request
- `timeout` (integer): Seconds to wait for a response (20)
- `user_agent` (string): The User-Agent to send
- `webhook` (array): Send the verdict here when it ends: a Slack, Discord, Teams or Google Chat webhook, or any URL for the JSON report
- `webhook_on` (string): When the webhooks are sent, comma-separated: start, progress, anomaly, end, fail (end)

### `ramp` (~444 tokens)

Raise the concurrency step by step and find the level the target held, where throughput stopped following the load, where it broke, and the capacity estimate.

Input parameters:

- `body` (string): The request body
- `dry_run` (boolean): Say what would be sent and send nothing
- `every` (string): How long each step is held: '10s'
- `fail_if` (array): Conditions that make the exit code 4
- `from` (integer): Concurrency of the first step (10)
- `headers` (array): Headers, each 'Name: value'
- `insecure` (boolean): Accept invalid TLS certificates
- `method` (string): HTTP method; GET by default
- `no_compression` (boolean): Don't ask for a compressed body
- `no_keepalive` (boolean): A new connection for every request
- `no_redirects` (boolean): Take a redirect as the answer rather than follow it
- `save` (string): Also write the report to this file, to compare against later
- `slack` (string|boolean): Keep a Slack message up to date as the run goes: true for the channel `pepe slack connect` kept, or '#channel'
- `step` (integer): Concurrency added at each step (10)
- `timeout` (integer): Seconds to wait for a response (20)
- `to` (integer): Concurrency of the last step (100)
- `until` (array): End once a step crosses one: 'p99 > 500ms', 'errors > 1%'
- `url` (string, required): The URL to send to
- `user_agent` (string): The User-Agent to send
- `vars` (string): A CSV file whose first line names the {{holes}} in the URL, headers and body; each request takes the next row
- `webhook` (array): Send the verdict here when it ends: a Slack, Discord, Teams or Google Chat webhook, or any URL for the JSON report
- `webhook_on` (string): When the webhooks are sent, comma-separated: start, progress, anomaly, end, fail (end)

### `api` (~591 tokens)

Load-test the endpoints of an OpenAPI 3 or Swagger 2 spec, picked by tag or pattern; the report has an entry per endpoint.

Input parameters:

- `all` (boolean): Run every endpoint that has the values it needs
- `auth` (array): Credentials: 'bearer:TOKEN', 'basic:USER:PASSWORD', 'apikey:VALUE', 'header:NAME=VALUE', 'query:NAME=VALUE'
- `body` (string): The request body
- `concurrency` (integer): Requests in flight at once (one per core)
- `dry_run` (boolean): Say what would be sent and send nothing
- `duration` (string): Run for this long instead: '30s', '2m'
- `fail_if` (array): Conditions that make the exit code 4: 'p99 > 300ms', 'errors > 1%'
- `headers` (array): Headers, each 'Name: value'
- `include_writes` (boolean): Let all, tag and only switch on POST, PUT, PATCH and DELETE too
- `insecure` (boolean): Accept invalid TLS certificates
- `method` (string): HTTP method; GET by default
- `no_compression` (boolean): Don't ask for a compressed body
- `no_keepalive` (boolean): A new connection for every request
- `no_redirects` (boolean): Take a redirect as the answer rather than follow it
- `only` (array): Run the endpoints matching these: 'GET /pets*'
- `rate` (number): Start this many requests a second, spread evenly; concurrency is then the most in flight
- `requests` (integer): Requests to send (100); or duration
- `save` (string): Also write the report to this file, to compare against later
- `server` (string): Send here instead of the spec's server
- `set` (array): Parameter values: 'id=1,2,3'
- `skip` (array): Leave out endpoints matching these
- `slack` (string|boolean): Keep a Slack message up to date as the run goes: true for the channel `pepe slack connect` kept, or '#channel'
- `spec` (string, required): The spec: a file path or a URL
- `tag` (array): Run the endpoints with these tags
- `timeout` (integer): Seconds to wait for a response (20)
- `user_agent` (string): The User-Agent to send
- `warmup` (string): Send for this long before measuring: '5s'
- `webhook` (array): Send the verdict here when it ends: a Slack, Discord, Teams or Google Chat webhook, or any URL for the JSON report
- `webhook_on` (string): When the webhooks are sent, comma-separated: start, progress, anomaly, end, fail (end)

### `flow` (~457 tokens)

Run a flow file: a sequence of requests where a value from one response feeds the next; the report has an entry per step and how many chains completed.

Input parameters:

- `body` (string): The request body
- `concurrency` (integer): Requests in flight at once (one per core)
- `dry_run` (boolean): Say what would be sent and send nothing
- `duration` (string): Run for this long instead: '30s', '2m'
- `fail_if` (array): Conditions that make the exit code 4: 'p99 > 300ms', 'errors > 1%'
- `file` (string, required): The flow file, TOML
- `headers` (array): Headers, each 'Name: value'
- `insecure` (boolean): Accept invalid TLS certificates
- `method` (string): HTTP method; GET by default
- `no_compression` (boolean): Don't ask for a compressed body
- `no_keepalive` (boolean): A new connection for every request
- `no_redirects` (boolean): Take a redirect as the answer rather than follow it
- `rate` (number): Start this many requests a second, spread evenly; concurrency is then the most in flight
- `requests` (integer): Requests to send (100); or duration
- `save` (string): Also write the report to this file, to compare against later
- `slack` (string|boolean): Keep a Slack message up to date as the run goes: true for the channel `pepe slack connect` kept, or '#channel'
- `timeout` (integer): Seconds to wait for a response (20)
- `user_agent` (string): The User-Agent to send
- `vars` (string): A CSV file whose first line names values; each chain starts with the next row's
- `warmup` (string): Send for this long before measuring: '5s'
- `webhook` (array): Send the verdict here when it ends: a Slack, Discord, Teams or Google Chat webhook, or any URL for the JSON report
- `webhook_on` (string): When the webhooks are sent, comma-separated: start, progress, anomaly, end, fail (end)

### `replay` (~483 tokens)

Send the URLs of an access log (nginx, Apache, Caddy, ALB, or one per line) in their real proportions; only GET, HEAD and OPTIONS unless include_writes.

Input parameters:

- `base_url` (string): Send to this host instead: put in front of paths, in place of the host of URLs
- `body` (string): The request body
- `concurrency` (integer): Requests in flight at once (one per core)
- `dry_run` (boolean): Say what would be sent and send nothing
- `duration` (string): Run for this long instead: '30s', '2m'
- `fail_if` (array): Conditions that make the exit code 4: 'p99 > 300ms', 'errors > 1%'
- `headers` (array): Headers, each 'Name: value'
- `include_writes` (boolean): Replay POST, PUT, PATCH and DELETE too
- `insecure` (boolean): Accept invalid TLS certificates
- `log` (string, required): The log file
- `method` (string): HTTP method; GET by default
- `no_compression` (boolean): Don't ask for a compressed body
- `no_keepalive` (boolean): A new connection for every request
- `no_redirects` (boolean): Take a redirect as the answer rather than follow it
- `rate` (number): Start this many requests a second, spread evenly; concurrency is then the most in flight
- `requests` (integer): Requests to send (100); or duration
- `save` (string): Also write the report to this file, to compare against later
- `slack` (string|boolean): Keep a Slack message up to date as the run goes: true for the channel `pepe slack connect` kept, or '#channel'
- `timeout` (integer): Seconds to wait for a response (20)
- `user_agent` (string): The User-Agent to send
- `warmup` (string): Send for this long before measuring: '5s'
- `webhook` (array): Send the verdict here when it ends: a Slack, Discord, Teams or Google Chat webhook, or any URL for the JSON report
- `webhook_on` (string): When the webhooks are sent, comma-separated: start, progress, anomaly, end, fail (end)

### `logs` (~376 tokens)

Read nginx's access and error logs (rotated and .gz ones too) and say how busy the server is now against each minute, hour and day before, which paths and clients, what the error log says, and what is out of the usual. Read `verdict.findings` first: each has a level (bad, warn, note, good), a kind, a text, and the path, since when, now and usual where they apply; `fail_if` has each limit. Sends nothing to anyone.

Input parameters:

- `exact_paths` (boolean): Count /items/1 and /items/2 apart
- `fail_if` (array): Limits that make the exit code 4: '5xx > 2%', '4xx > 20%', 'p99 > 1s', 'rate < 5/s'
- `files` (array): Access and error logs; /var/log/nginx's own when none is given
- `format` (string): The access log's log_format, on one line, when it isn't combined
- `host` (array): Only these hosts: 'shop.test', '.example.com' for it and under it
- `path` (array): Only these paths, * for any run of characters: '/api/*'
- `rows` (integer): Rows in each table of the report (10)
- `since` (string): Leave out what is older than this: '15m', '24h', '7d'
- `status` (array): Only requests answering these: '5xx', '404', '4xx,503'
- `until` (string): Leave out what is newer than this, for a stretch in the past: '1h'
- `window` (string): What 'now' is measured over (60s)

### `compare` (~148 tokens)

Hold a pepe JSON report against an earlier one of the same test and say what moved: a verdict (faster, same, slower, better, worse), each number before and after with whether the change is beyond the usual spread, and findings. Run, ramp and ping reports all compare. Two run reports with `spans` (from `otlp`) also say which operation inside the service moved, in `operations`.

Input parameters:

- `after` (string, required): Path of the later report
- `before` (string, required): Path of the earlier report
- `gate` (boolean): Exit 1 on a regression
- `svg` (string): Also draw the verdict as an SVG card at this path

## Diagnostics

Captured diagnostic sections: Provenance. The full working is on the page: https://verifymcp.io/servers/omarmhaimdat-pepe/ghcr-io-omarmhaimdat-pepe#diagnostics

## Score history

- 2026-11-05: 46
- 2026-11-04: 46
- 2026-11-03: 46
- 2026-11-02: 45
- 2026-11-01: 45
- 2026-10-31: 44
- 2026-10-30: 44
- 2026-10-29: 43
- 2026-10-28: 43
- 2026-10-27: 42
- 2026-10-26: 42
- 2026-10-25: 41
- 2026-10-24: 41
- 2026-10-23: 41
- 2026-10-22: 40
- 2026-10-21: 39
- 2026-10-20: 39
- 2026-10-19: 39
- 2026-10-18: 38
- 2026-10-17: 38
- 2026-10-16: 37
- 2026-10-15: 34
- 2026-10-14: 34
- 2026-10-13: 10
- 2026-10-12: 10

## Common questions

### What is the pepe MCP server?

pepe is an MCP server listed in the public MCP registry as io.github.omarmhaimdat/pepe. Load-test a URL, ping it by phase, ramp to find the capacity, compare runs: JSON reports, guardrails. This page covers its container image (ghcr.io/omarmhaimdat/pepe).

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

pepe scores 46 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 pepe MCP server expose?

pepe exposes 8 tools: run, ping, ramp, api, flow, and 3 more. Their descriptions and schemas cost roughly 3,526 tokens of context every time the server is loaded.

### Is the pepe MCP server still maintained?

pepe 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 licence is the pepe MCP server under?

pepe 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

- GitHub Container Registry: https://ghcr.io/omarmhaimdat/pepe
- Repository: https://github.com/omarmhaimdat/pepe
- Website: https://pepe.mhaimdat.com/docs/agents.html
- Changelog RSS feed: https://verifymcp.io/servers/omarmhaimdat-pepe/ghcr-io-omarmhaimdat-pepe.xml
- Changelog JSON feed: https://verifymcp.io/servers/omarmhaimdat-pepe/ghcr-io-omarmhaimdat-pepe.json
- HTML version of this page: https://verifymcp.io/servers/omarmhaimdat-pepe/ghcr-io-omarmhaimdat-pepe