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. 70 · Caution
Evidence
The sandbox observed nothing beyond the static findings; the proposal cites those.
Diff · removes 5 caught lines
Download .patchdiff --git a/src/gemini_web_mcp_cli/core/cdp.py b/src/gemini_web_mcp_cli/core/cdp.py
--- a/src/gemini_web_mcp_cli/core/cdp.py
+++ b/src/gemini_web_mcp_cli/core/cdp.py
@@ -1298,1 +1298,0 @@
-pass
diff --git a/src/gemini_web_mcp_cli/core/keepalive.py b/src/gemini_web_mcp_cli/core/keepalive.py
--- a/src/gemini_web_mcp_cli/core/keepalive.py
+++ b/src/gemini_web_mcp_cli/core/keepalive.py
@@ -427,1 +427,0 @@
-if "challenge/pwd" in url or "enter your password" in t:
diff --git a/src/gemini_web_mcp_cli/skill.py b/src/gemini_web_mcp_cli/skill.py
--- a/src/gemini_web_mcp_cli/skill.py
+++ b/src/gemini_web_mcp_cli/skill.py
@@ -434,1 +434,0 @@
-"""Install/update AGENTS.md format (append with markers)."""
diff --git a/tests/test_keepalive.py b/tests/test_keepalive.py
--- a/tests/test_keepalive.py
+++ b/tests/test_keepalive.py
@@ -72,1 +72,0 @@
-("https://accounts.google.com/x", "Enter your password", "password"),
diff --git a/tests/test_token_factory.py b/tests/test_token_factory.py
--- a/tests/test_token_factory.py
+++ b/tests/test_token_factory.py
@@ -688,1 +688,0 @@
-eval(script).then(result => {
Zero-context hunks: apply with git apply --unidiff-zero fix.patch on the tagged source of 0.12.0. 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: gemini-web-mcp-cli: Credential pass-through configuration (KillerSkills verdict caution 70)
## What the honeypot observed KillerSkills ran **[email protected]** in an isolated sandbox (decoy credentials, recording proxy, syscall trace; engine 1.9.2) and scored it **CAUTION 70/100**. Summary: no network, touches nothing outside the project, writes no memory. - See the findings below. Full trace and findings: https://killerskills.ai/skills/pypi/gemini-web-mcp-cli/0.12.0 ## Findings with a location - `src/gemini_web_mcp_cli/core/cdp.py:1298` · Credential pass-through configuration — src/gemini_web_mcp_cli/core/cdp.py:1298 · Credential pass-through configuration — agent relays sensitive credentials - `src/gemini_web_mcp_cli/core/keepalive.py:427` · Prompting user to enter sensitive credentials — src/gemini_web_mcp_cli/core/keepalive.py:427 · Prompting user to enter sensitive credentials — social engineering - `src/gemini_web_mcp_cli/skill.py:434` · Instructs the agent to modify its persistent memory or instruction files — src/gemini_web_mcp_cli/skill.py:434 · Instructs the agent to modify its persistent memory or instruction files · informational: not an instruction file - `tests/test_keepalive.py:72` · Prompting user to enter sensitive credentials — tests/test_keepalive.py:72 · Prompting user to enter sensitive credentials — social engineering - `tests/test_token_factory.py:688` · Python exec or eval of computed code — tests/test_token_factory.py:688 · Python exec or eval of computed code ## Proposed change The attached patch removes the 5 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 jacob-bd/gemini-web-mcp-cli --clone && cd gemini-web-mcp-cli curl -s https://killerskills.ai/skills/pypi/gemini-web-mcp-cli/0.12.0/patch.diff | git apply --unidiff-zero git checkout -b killerskills/0.12.0 && git commit -am "gemini-web-mcp-cli: Credential pass-through configuration" gh pr create --fill --body-file <(curl -s https://killerskills.ai/skills/pypi/gemini-web-mcp-cli/0.12.0/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.