Sealmark
Automated re-test on 2026-11-21: not checked by hand, and not yet reviewed by the maintainers.

Results / @pidgeon-ai/mcp

@pidgeon-ai/mcp

version tested 1.15.0npmtested 2026-11-21Sealmark 1.0.1methodology 0.3-draftSTANDARD run84 tools

Sealmark's automated test of @pidgeon-ai/mcp 1.15.0 on 2026-11-21 gave it grade D, its lowest test-area grade (protocol D and effects A). Only 16 of 84 tools could be called with default settings, so the deeper test areas cover only those tools. Of its 10 checks, 5 passed, 1 failed, 2 inconclusive and 2 not applicable. 3 more checks are advisory, and advisory checks never change a grade: 1 passed and 2 not applicable. The run used Sealmark 1.0.1 with the standard settings. The grade includes a finding under private disclosure; it still counts in full.

D

Grade D

The server's grade is its lowest test-area grade. How grading works

The grade includes a finding under private disclosure. It still counts in full.

Only 16 of 84 tools could be called with default settings, so the deeper test areas cover only those tools.

Grade by test area

D
Protocol3/4 points · capped at Dcritical: tool-poisoningincludes a finding under private disclosure
A
Effects2/2 points

Protocol checks

CheckVerdictWhat happened
protocol
Valid MCP handshake and tool schemas
pass
handshake ok, 84 well-formed tool(s)
tool-poisoningcritical
No tool-poisoning indicators in tool metadata
fail
A security finding is being reported to the maintainer privately. Details will be published when it is fixed or 90 days after the report (disclosure policy).
robust-input
Survives malformed tool input
pass
112 invalid call(s) rejected cleanly
secret-echocritical
No environment secret echoed in responses
pass
Details held back while a security finding in this test area is reported privately.
instructions-hygiene (advisory)
Server instructions name only tools that exist
pass
the instructions name no tool the server lacks
refusal-quality (advisory)
Refusals say what to do next
n/a
no refusal was recorded in this run, so there is nothing to judge

Network checks

CheckVerdictWhat happened
net-loopback-ssrfcritical
Refuses loopback targets (SSRF)
inconclusive
the listener saw nothing, but 9 of 9 probes were not explicitly refused (9 failed without saying the target was refused), so a refusal cannot be told from a tool that cannot fetch at all (a reply that only reports a failed fetch (proxy, socket, DNS) is not read as a refusal; set net.control_url to a URL the tool can reach)
net-scheme-allowlistcritical
Only http(s) schemes accepted
inconclusive
no scheme had an observable effect, but 6 of 6 scored probes were not explicitly refused, so a refusal cannot be told from a tool that cannot fetch at all (a reply that only reports a failed fetch (proxy, socket, DNS) is not read as a refusal; set net.control_url to a URL the tool can reach)
net-caps
Response size and time are capped
n/a
net-loopback-ssrf was inconclusive: the server never reached the loopback listener, so no loopback URL is known to be fetchable; response size and time caps were not measured
net-redirect
Redirects are re-validated
n/a
the server refused or never reached the loopback listener, so redirect handling could not be observed (a server that blocks loopback cannot be sent a redirect it would follow)

Effects checks

CheckVerdictWhat happened
annotation-honestycritical
Read-only tools change nothing
pass
status declared read-only and changed nothing in the observed directories during a benign call (list_inboxes, list_threads, find_contact, list_rules, list_suppressions, get_storage, list_domains, list_pods, list_events, list_api_keys, list_webhooks, list_agents, get_reputation, list_sites, list_mandates not judged: the control call failed)
openworld-honesty
Closed-world tools stay on the machine
pass
16 closed-world tool(s) called; nothing connected to the loopback listener (path and git tools cannot be pointed at it, so this confirms only that nothing connected out during the call)
undeclared-writes (advisory)
Unannotated tools leave the sandbox unchanged
n/a
every tool carries annotations, nothing to check

Each check id links to what it tests and how to fix a failure. The same text is in sealmark explain <id>.

What was not covered

Grade over time

Since the run on

  • Server version changed from 1.14.0 to 1.15.0.
  • The server listed 84 tools, 1 more than before (83).

Unchanged: the same grade in every test area, the same Sealmark version (1.0.1).

  1. Av1.3.0Sealmark 1.0.1
  2. Dv1.6.0Sealmark 1.0.1
  3. Dv1.7.0Sealmark 1.0.1
  4. Dv1.9.1Sealmark 1.0.1
  5. Dv1.11.0Sealmark 1.0.1
  6. Dv1.14.0Sealmark 1.0.1
  7. Dv1.15.0Sealmark 1.0.1

Package facts

License
Apache-2.0

From npm, checked 22 Nov 2026. Plain facts about the package; they play no part in the grade.

Run it yourself

The command for this run is withheld while a security finding is reported privately. It is published with the report, when the finding is fixed or 90 days after the report.

Grade alerts

One email when the grade of @pidgeon-ai/mcp changes after a re-test. Nothing else, and every email has a link to stop them.

Each of 613 verdicts was checked independently, and Sealmark was wrong on some: how reliable Sealmark is.

Is this wrong?

Reply to our private report or email disclosure@sealmark.dev and we will recheck it. Maintainers can ask for their server to be removed.

If we were wrong, we correct the page and add your reply. Security findings go to the maintainer privately first; see the disclosure policy.