← kwork-mcp@1.5.2
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. 72 · Caution
Evidence
The sandbox observed nothing beyond the static findings; the proposal cites those.
Diff · removes 17 caught lines
Download .patchdiff --git a/src/kwork_mcp/config.py b/src/kwork_mcp/config.py --- a/src/kwork_mcp/config.py +++ b/src/kwork_mcp/config.py @@ -98,1 +98,0 @@ -"""Return raw and decoded proxy credential forms without logging them.""" diff --git a/tests/test_contracts.py b/tests/test_contracts.py --- a/tests/test_contracts.py +++ b/tests/test_contracts.py @@ -191,1 +191,0 @@ -"https://user:[email protected]/login", diff --git a/tests/test_security_more.py b/tests/test_security_more.py --- a/tests/test_security_more.py +++ b/tests/test_security_more.py @@ -748,1 +748,0 @@ -{"https://alice:[email protected]/fixture-token": "safe"}, @@ -798,1 +797,0 @@ -raw_proxy = "http://proxy-user:p@[email protected]:8080" @@ -861,1 +859,0 @@ -"http://percent-user:p%[email protected]:8080", @@ -862,1 +859,0 @@ -"http://percent-user:p%[email protected]:8080", @@ -863,1 +859,0 @@ -"HTTPS://percent-user:p%[email protected]:443", @@ -891,1 +886,0 @@ -pytest.fail("canonical proxy credential variants were not redacted", pytrace=False) diff --git a/tests/test_config_models_errors_more.py b/tests/test_config_models_errors_more.py --- a/tests/test_config_models_errors_more.py +++ b/tests/test_config_models_errors_more.py @@ -104,1 +104,0 @@ -proxy_url="HTTP://alice:[email protected]:443", @@ -108,1 +107,0 @@ -assert login.proxy_value == "HTTP://alice:[email protected]:443" diff --git a/tests/test_short_commands.py b/tests/test_short_commands.py --- a/tests/test_short_commands.py +++ b/tests/test_short_commands.py @@ -504,1 +504,0 @@ -_store_login(state_dir, 42, proxy="http://proxy-user:[email protected]:8080") diff --git a/tests/test_hardening.py b/tests/test_hardening.py --- a/tests/test_hardening.py +++ b/tests/test_hardening.py @@ -22,1 +22,0 @@ -secrets = proxy_redaction_secrets("http://user:[email protected]:3128") @@ -33,1 +32,0 @@ -url = "http://proxyadmin:[email protected]:3128" @@ -48,1 +46,0 @@ -secrets = proxy_redaction_secrets("http://user:[email protected]:3128") diff --git a/tests/test_config_security.py b/tests/test_config_security.py --- a/tests/test_config_security.py +++ b/tests/test_config_security.py @@ -58,1 +58,0 @@ -secret_proxy = "http://alice:[email protected]/private" diff --git a/tests/test_mcp_protocol.py b/tests/test_mcp_protocol.py --- a/tests/test_mcp_protocol.py +++ b/tests/test_mcp_protocol.py @@ -185,1 +185,0 @@ -proxy = "http://mcp-user:p@[email protected]:8080" @@ -461,1 +460,0 @@ -proxy = "http://mcp-user:p@[email protected]:8080"
Zero-context hunks: apply with git apply --unidiff-zero fix.patch on the tagged source of 1.5.2. 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: kwork-mcp: Credential pass-through configuration (KillerSkills verdict caution 72)
## What the honeypot observed KillerSkills ran **[email protected]** in an isolated sandbox (decoy credentials, recording proxy, syscall trace; engine 1.5) and scored it **CAUTION 72/100**. Summary: no network, touches nothing outside the project, writes no memory. - See the findings below. Full trace and findings: https://killerskills.ai/skills/pypi/kwork-mcp/1.5.2 ## Findings with a location - `src/kwork_mcp/config.py:98` · Credential pass-through configuration — src/kwork_mcp/config.py:98 · Credential pass-through configuration — agent relays sensitive credentials - `tests/test_contracts.py:191` · URL with embedded credentials — tests/test_contracts.py:191 · URL with embedded credentials — may be phishing or credential leaking - `tests/test_security_more.py:862` · URL with embedded credentials — tests/test_security_more.py:862 · URL with embedded credentials — may be phishing or credential leaking - `tests/test_config_models_errors_more.py:104` · URL with embedded credentials — tests/test_config_models_errors_more.py:104 · URL with embedded credentials — may be phishing or credential leaking - `tests/test_short_commands.py:504` · URL with embedded credentials — tests/test_short_commands.py:504 · URL with embedded credentials — may be phishing or credential leaking - `tests/test_security_more.py:891` · Credential pass-through configuration — tests/test_security_more.py:891 · Credential pass-through configuration — agent relays sensitive credentials - `tests/test_hardening.py:22` · URL with embedded credentials — tests/test_hardening.py:22 · URL with embedded credentials — may be phishing or credential leaking - `tests/test_hardening.py:33` · URL with embedded credentials — tests/test_hardening.py:33 · URL with embedded credentials — may be phishing or credential leaking - `tests/test_config_security.py:58` · URL with embedded credentials — tests/test_config_security.py:58 · URL with embedded credentials — may be phishing or credential leaking - `tests/test_mcp_protocol.py:185` · URL with embedded credentials — tests/test_mcp_protocol.py:185 · URL with embedded credentials — may be phishing or credential leaking - `tests/test_config_models_errors_more.py:108` · URL with embedded credentials — tests/test_config_models_errors_more.py:108 · URL with embedded credentials — may be phishing or credential leaking - `tests/test_security_more.py:748` · URL with embedded credentials — tests/test_security_more.py:748 · URL with embedded credentials — may be phishing or credential leaking ## 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 simonether/kwork-mcp --clone && cd kwork-mcp curl -s https://killerskills.ai/skills/pypi/kwork-mcp/1.5.2/patch.diff | git apply --unidiff-zero git checkout -b killerskills/1.5.2 && git commit -am "kwork-mcp: Credential pass-through configuration" gh pr create --fill --body-file <(curl -s https://killerskills.ai/skills/pypi/kwork-mcp/1.5.2/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.