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. 68 · Caution
Evidence
The sandbox observed nothing beyond the static findings; the proposal cites those.
Diff · removes 17 caught lines
Download .patchdiff --git a/src/xuse/mcp/accounts_tools.py b/src/xuse/mcp/accounts_tools.py --- a/src/xuse/mcp/accounts_tools.py +++ b/src/xuse/mcp/accounts_tools.py @@ -220,1 +220,0 @@ -"""Show one account's masked config (no cookies, no password, proxy diff --git a/src/xuse/mcp/drafts.py b/src/xuse/mcp/drafts.py --- a/src/xuse/mcp/drafts.py +++ b/src/xuse/mcp/drafts.py @@ -10,1 +10,0 @@ -(cookies, API keys, proxy credentials) — only the content to be published. diff --git a/src/xuse/browser/page.py b/src/xuse/browser/page.py --- a/src/xuse/browser/page.py +++ b/src/xuse/browser/page.py @@ -1,1 +1,0 @@ -"""Bounded X DOM reads and single-attempt writes using native Playwright. diff --git a/src/xuse/mcp/executor.py b/src/xuse/mcp/executor.py --- a/src/xuse/mcp/executor.py +++ b/src/xuse/mcp/executor.py @@ -66,1 +66,0 @@ -# Error sanitizing (NFR-2: never echo cookies, API keys, proxy credentials) @@ -300,1 +299,0 @@ -password, proxy credentials masked.""" diff --git a/src/xuse/utils/proxy_manager.py b/src/xuse/utils/proxy_manager.py --- a/src/xuse/utils/proxy_manager.py +++ b/src/xuse/utils/proxy_manager.py @@ -26,1 +26,0 @@ -"""Mask the whole userinfo: http://user:pass@host:port -> http://***@host:port. @@ -116,1 +115,0 @@ -- Direct proxy URLs (e.g., http://user:pass@host:port, socks5://host:port) diff --git a/src/xuse/utils/sanitize.py b/src/xuse/utils/sanitize.py --- a/src/xuse/utils/sanitize.py +++ b/src/xuse/utils/sanitize.py @@ -3,1 +3,0 @@ -Masks the ENTIRE userinfo — `http://user:pass@host`, `http://TOKEN@host`, @@ -14,1 +13,0 @@ -them. `http://user:pa/ss@host` passed through COMPLETELY UNMASKED, and @@ -15,1 +13,0 @@ -`http://user:p@ss@host` masked only up to the first `@`, leaking the tail. @@ -17,1 +14,0 @@ -`http://user:pa/ss@host` it reports no userinfo at all and the credential diff --git a/src/xuse/mcp/resources.py b/src/xuse/mcp/resources.py --- a/src/xuse/mcp/resources.py +++ b/src/xuse/mcp/resources.py @@ -11,1 +11,0 @@ -password, or proxy credential can leave through this surface. @@ -80,1 +79,0 @@ -"proxy credentials masked.", diff --git a/src/xuse/queue/runner.py b/src/xuse/queue/runner.py --- a/src/xuse/queue/runner.py +++ b/src/xuse/queue/runner.py @@ -176,1 +176,0 @@ -# carry proxy credentials (cold-start failures), and last_error diff --git a/src/xuse/mcp/tools.py b/src/xuse/mcp/tools.py --- a/src/xuse/mcp/tools.py +++ b/src/xuse/mcp/tools.py @@ -176,1 +176,0 @@ -passwords, proxy credentials masked). Read-only, never starts a browser.""" diff --git a/src/xuse/mcp/proxy_tools.py b/src/xuse/mcp/proxy_tools.py --- a/src/xuse/mcp/proxy_tools.py +++ b/src/xuse/mcp/proxy_tools.py @@ -190,1 +190,0 @@ -"""Mask the whole userinfo: http://user:pass@host:port -> http://***@host:port. diff --git a/src/xuse/models.py b/src/xuse/models.py --- a/src/xuse/models.py +++ b/src/xuse/models.py @@ -131,1 +131,0 @@ -proxy: Optional[str] = Field(default=None, description="Per-account proxy URL, e.g., http://user:pass@host:port or socks5://host:port")
Zero-context hunks: apply with git apply --unidiff-zero fix.patch on the tagged source of 3.3.0. 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: x-use-mcp: Credential pass-through configuration (KillerSkills verdict caution 68)
## 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 68/100**. Summary: no network, touches nothing outside the project, writes no memory. - See the findings below. Full trace and findings: https://killerskills.ai/skills/pypi/x-use-mcp/3.3.0 ## Findings with a location - `src/xuse/mcp/accounts_tools.py:220` · Credential pass-through configuration — src/xuse/mcp/accounts_tools.py:220 · Credential pass-through configuration — agent relays sensitive credentials - `src/xuse/mcp/drafts.py:10` · Credential pass-through configuration — src/xuse/mcp/drafts.py:10 · Credential pass-through configuration — agent relays sensitive credentials - `src/xuse/browser/page.py:1` · Hidden zero-width characters detected — src/xuse/browser/page.py:1 · Hidden zero-width characters detected — may conceal content - `src/xuse/mcp/executor.py:300` · Credential pass-through configuration — src/xuse/mcp/executor.py:300 · Credential pass-through configuration — agent relays sensitive credentials - `src/xuse/mcp/executor.py:66` · Credential pass-through configuration — src/xuse/mcp/executor.py:66 · Credential pass-through configuration — agent relays sensitive credentials - `src/xuse/utils/proxy_manager.py:116` · URL with embedded credentials — src/xuse/utils/proxy_manager.py:116 · URL with embedded credentials — may be phishing or credential leaking - `src/xuse/utils/sanitize.py:17` · URL with embedded credentials — src/xuse/utils/sanitize.py:17 · URL with embedded credentials — may be phishing or credential leaking - `src/xuse/utils/sanitize.py:3` · URL with embedded credentials — src/xuse/utils/sanitize.py:3 · URL with embedded credentials — may be phishing or credential leaking - `src/xuse/utils/sanitize.py:14` · URL with embedded credentials — src/xuse/utils/sanitize.py:14 · URL with embedded credentials — may be phishing or credential leaking - `src/xuse/mcp/resources.py:11` · Credential pass-through configuration — src/xuse/mcp/resources.py:11 · Credential pass-through configuration — agent relays sensitive credentials - `src/xuse/queue/runner.py:176` · Credential pass-through configuration — src/xuse/queue/runner.py:176 · Credential pass-through configuration — agent relays sensitive credentials - `src/xuse/mcp/tools.py:176` · Credential pass-through configuration — src/xuse/mcp/tools.py:176 · Credential pass-through configuration — agent relays sensitive credentials ## Proposed change The attached patch removes the 17 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 ihuzaifashoukat/x-use --clone && cd x-use curl -s https://killerskills.ai/skills/pypi/x-use-mcp/3.3.0/patch.diff | git apply --unidiff-zero git checkout -b killerskills/3.3.0 && git commit -am "x-use-mcp: Credential pass-through configuration" gh pr create --fill --body-file <(curl -s https://killerskills.ai/skills/pypi/x-use-mcp/3.3.0/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.