← mcp-server-git@2026.11.4

Propose a patch

A fix the maintainer can act on, with the evidence attached. It opens under your GitHub account in your browser; KillerSkills holds no token and never modifies the package. The fix is verified the only way we verify anything: the next crawl re-runs the new version in the honeypot. 74 · Caution

Evidence

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

Diff · removes 8 caught lines

Download .patch
diff --git a/src/mcp_server_git/server.py b/src/mcp_server_git/server.py
--- a/src/mcp_server_git/server.py
+++ b/src/mcp_server_git/server.py
@@ -682,1 +682,0 @@
-# (`https::https://user:token@host/...`, gitremote-helpers(7)). The address
@@ -705,1 +704,0 @@
-A remote URL can carry a credential (`https://user:token@host/...`), and
diff --git a/README.md b/README.md
--- a/README.md
+++ b/README.md
@@ -232,1 +232,0 @@
-<summary>Using docker</summary> * Note: replace '/Users/username' with a path that you want to be accessible by this tool * Note: the `-e` lines set the identit
diff --git a/tests/test_remote.py b/tests/test_remote.py
--- a/tests/test_remote.py
+++ b/tests/test_remote.py
@@ -4,1 +4,0 @@
-(`https://user:token@host/...`), so most of these pin `redact_url`'s rule: the
@@ -49,1 +48,0 @@
-repo.create_remote("origin", "https://user:[email protected]/org/repo.git")
@@ -90,1 +88,0 @@
-"https://user:[email protected]/r.git",
@@ -94,1 +91,0 @@
-("http://u:[email protected]:8080/r", "http://<redacted>@example.com:8080/r"),
@@ -95,1 +91,0 @@
-("HTTPS://u:[email protected]/r", "HTTPS://<redacted>@example.com/r"),

Zero-context hunks: apply with git apply --unidiff-zero fix.patch on the tagged source of 2026.11.4. 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: mcp-server-git: URL with embedded credentials (KillerSkills verdict caution 74)

## 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 74/100**. Summary: no network, touches nothing outside the project, writes no memory.

- See the findings below.

Full trace and findings: https://killerskills.ai/skills/pypi/mcp-server-git/2026.11.4

## Findings with a location
- `src/mcp_server_git/server.py:682` · URL with embedded credentials — src/mcp_server_git/server.py:682 · URL with embedded credentials — may be phishing or credential leaking
- `README.md:232` · Collapsed <details> block carrying instructions — README.md:232 · Collapsed <details> block carrying instructions
- `tests/test_remote.py:95` · URL with embedded credentials — tests/test_remote.py:95 · URL with embedded credentials — may be phishing or credential leaking
- `src/mcp_server_git/server.py:705` · URL with embedded credentials — src/mcp_server_git/server.py:705 · URL with embedded credentials — may be phishing or credential leaking
- `tests/test_remote.py:4` · URL with embedded credentials — tests/test_remote.py:4 · URL with embedded credentials — may be phishing or credential leaking
- `tests/test_remote.py:94` · URL with embedded credentials — tests/test_remote.py:94 · URL with embedded credentials — may be phishing or credential leaking
- `tests/test_remote.py:90` · URL with embedded credentials — tests/test_remote.py:90 · URL with embedded credentials — may be phishing or credential leaking
- `tests/test_remote.py:49` · URL with embedded credentials — tests/test_remote.py:49 · URL with embedded credentials — may be phishing or credential leaking

## Proposed change
The attached patch removes the 8 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._
# or from a terminal, under your own account:
gh repo fork modelcontextprotocol/servers --clone && cd servers
curl -s https://killerskills.ai/skills/pypi/mcp-server-git/2026.11.4/patch.diff | git apply --unidiff-zero
git checkout -b killerskills/2026.11.4 && git commit -am "mcp-server-git: URL with embedded credentials"
gh pr create --fill --body-file <(curl -s https://killerskills.ai/skills/pypi/mcp-server-git/2026.11.4/patch.md)

We count proposals opened, not merged: a merge shows up as a new version that the crawl re-runs. Disputes: the maintainer can contest the verdict from its page; a person reviews it and the outcome is published.