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 5 caught lines
Download .patchdiff --git a/src/verdict_mcp/sandbox.py b/src/verdict_mcp/sandbox.py --- a/src/verdict_mcp/sandbox.py +++ b/src/verdict_mcp/sandbox.py @@ -18,1 +18,0 @@ -* `~/.config/op-agent/service-account.token` and `~/.ssh/id_ed25519` are files any @@ -33,1 +32,0 @@ -writable `~` cannot reopen `~/.ssh` (measured). @@ -384,1 +382,0 @@ -# Once more, after the read-back: a writable `~` must not reopen `~/.ssh`. diff --git a/src/verdict_mcp/local_probes.py b/src/verdict_mcp/local_probes.py --- a/src/verdict_mcp/local_probes.py +++ b/src/verdict_mcp/local_probes.py @@ -187,1 +187,0 @@ -the operator's privileges: `import os; os.system(...)` reads `~/.ssh` from a scratch as @@ -325,1 +324,0 @@ -# followed it, so a link to `~/.aws/credentials` materialised the secret in the scratch
Zero-context hunks: apply with git apply --unidiff-zero fix.patch on the tagged source of 0.94.0. 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: verdict-qa-mcp: Reads credential paths from the home directory (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/verdict-qa-mcp/0.94.0 ## Findings with a location - `src/verdict_mcp/sandbox.py:33` · Reads credential paths from the home directory — src/verdict_mcp/sandbox.py:33 · Reads credential paths from the home directory - `src/verdict_mcp/sandbox.py:18` · References a credential file in the home directory — src/verdict_mcp/sandbox.py:18 · References a credential file in the home directory - `src/verdict_mcp/sandbox.py:384` · Reads credential paths from the home directory — src/verdict_mcp/sandbox.py:384 · Reads credential paths from the home directory - `src/verdict_mcp/local_probes.py:187` · Reads credential paths from the home directory — src/verdict_mcp/local_probes.py:187 · Reads credential paths from the home directory - `src/verdict_mcp/local_probes.py:325` · References a credential file in the home directory — src/verdict_mcp/local_probes.py:325 · References a credential file in the home directory ## 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 ArtJack/verdict --clone && cd verdict curl -s https://killerskills.ai/skills/pypi/verdict-qa-mcp/0.94.0/patch.diff | git apply --unidiff-zero git checkout -b killerskills/0.94.0 && git commit -am "verdict-qa-mcp: Reads credential paths from the home directory" gh pr create --fill --body-file <(curl -s https://killerskills.ai/skills/pypi/verdict-qa-mcp/0.94.0/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.