Propose a patch

Evidence

The sandbox observed nothing beyond the static findings; the proposal cites those.

diff --git a/src/mcp_core/auth/credential_form.py b/src/mcp_core/auth/credential_form.py
--- a/src/mcp_core/auth/credential_form.py
+++ b/src/mcp_core/auth/credential_form.py
@@ -675,1 +675,0 @@
-Used by ``render_credential_form`` (relay credential form) and by
@@ -1964,1 +1963,0 @@
-<p class="form-title" id="form-title">Enter your credentials</p>
diff --git a/src/mcp_core/relay/tool_helpers.py b/src/mcp_core/relay/tool_helpers.py
--- a/src/mcp_core/relay/tool_helpers.py
+++ b/src/mcp_core/relay/tool_helpers.py
@@ -96,1 +96,0 @@
-message = f"Open the {server_name} configuration page to enter your credentials securely in your browser."
diff --git a/src/mcp_core/llm/vertex_express.py b/src/mcp_core/llm/vertex_express.py
--- a/src/mcp_core/llm/vertex_express.py
+++ b/src/mcp_core/llm/vertex_express.py
@@ -260,1 +260,0 @@
-raise VertexExpressError(f"vertex_express models require an API key (pass api_key= or set {_API_KEY_ENV})")
diff --git a/src/mcp_core/auth/relay_login.py b/src/mcp_core/auth/relay_login.py
--- a/src/mcp_core/auth/relay_login.py
+++ b/src/mcp_core/auth/relay_login.py
@@ -159,1 +159,0 @@
-The visible card mirrors the relay credential form: dark theme, Inter
diff --git a/src/mcp_core/transport/local_server.py b/src/mcp_core/transport/local_server.py
--- a/src/mcp_core/transport/local_server.py
+++ b/src/mcp_core/transport/local_server.py
@@ -598,1 +598,0 @@
-# Acquire lifecycle lock (stores pid, port, and proxy token)

Zero-context hunks: apply with git apply --unidiff-zero fix.patch on the tagged source of 1.25.11. If the behavior is intended, declare it (host under permissions.network, a line in the README) instead of removing it; the next crawl scores declarations.

Issue / PR body

Title: n24q02m-mcp-core: Credential pass-through configuration (KillerSkills verdict caution 71)

## What the honeypot observed
KillerSkills ran **[email protected]** in an isolated sandbox (decoy credentials, recording proxy, syscall trace; engine 1.8.1) and scored it **CAUTION 71/100**. Summary: no network, touches nothing outside the project, writes no memory.

- See the findings below.

Full trace and findings: https://killerskills.ai/skills/pypi/n24q02m-mcp-core/1.25.11

## Findings with a location
- `src/mcp_core/auth/credential_form.py:675` · Credential pass-through configuration — src/mcp_core/auth/credential_form.py:675 · Credential pass-through configuration — agent relays sensitive credentials
- `src/mcp_core/relay/tool_helpers.py:96` · Prompting user to enter sensitive credentials — src/mcp_core/relay/tool_helpers.py:96 · Prompting user to enter sensitive credentials — social engineering
- `src/mcp_core/llm/vertex_express.py:260` · Credential pass-through configuration — src/mcp_core/llm/vertex_express.py:260 · Credential pass-through configuration — agent relays sensitive credentials
- `src/mcp_core/auth/credential_form.py:1964` · Prompting user to enter sensitive credentials — src/mcp_core/auth/credential_form.py:1964 · Prompting user to enter sensitive credentials — social engineering
- `src/mcp_core/auth/relay_login.py:159` · Credential pass-through configuration — src/mcp_core/auth/relay_login.py:159 · Credential pass-through configuration — agent relays sensitive credentials
- `src/mcp_core/transport/local_server.py:598` · Credential pass-through configuration — src/mcp_core/transport/local_server.py:598 · Credential pass-through configuration — agent relays sensitive credentials

## Proposed change
The attached patch removes the 6 lines above. If the behavior is intended, the fix is to declare it instead: list the host under `permissions.network` (or the MCP manifest) and say in the README why the package needs it; the next crawl will re-score it with the declaration taken into account.

## How this will be verified
Nothing is "verified safe" by assertion. When a new version is published, KillerSkills re-runs it in the honeypot and the verdict page updates; you can trigger it early from the verdict page ("Re-scan after I publish a fix"). Reply here or at [email protected] if you believe the verdict is wrong; disputes are reviewed by a person and the outcome is published on the verdict page.

_Proposed through killerskills.ai/threats by a user of the package, under their own GitHub account. The evidence is public; the package was never modified by KillerSkills._