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. 71 · Caution

Evidence

Contacted localhost: not declared and not needed for the documented job.

Diff · removes 1 caught line

Download .patch

diff --git a/mcp:tools/agentistics_task_claim b/mcp:tools/agentistics_task_claim
--- a/mcp:tools/agentistics_task_claim
+++ b/mcp:tools/agentistics_task_claim
@@ -1,1 +1,0 @@
-erriding a live claim on purpose and an agent should not send it. {"type":"object","properties":{"ref":{"type":"string"},

Zero-context hunks: apply with git apply --unidiff-zero fix.patch on the tagged source of 2.205.0. 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: mcp: Contacted localhost: not declared and not needed for the documented job (KillerSkills verdict caution 71)

## 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 71/100**. Summary: sends data to localhost.

- Contacted localhost: not declared and not needed for the documented job.

Full trace and findings: https://killerskills.ai/skills/npm/mcp/2.205.0

## Findings with a location
- `mcp:tools/agentistics_task_claim:1` · Tool description instructs the agent: instruction addressed to the agent — mcp:tools/agentistics_task_claim:1 · Instruction aimed at the agent: instruction addressed to the agent

## Proposed change
The attached patch removes the 1 line 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._

Open the issue on GitHub (agentistics/agentistics)Fork to open a PR

# or from a terminal, under your own account:
gh repo fork agentistics/agentistics --clone && cd agentistics
curl -s https://killerskills.ai/skills/npm/mcp/2.205.0/patch.diff | git apply --unidiff-zero
git checkout -b killerskills/2.205.0 && git commit -am "mcp: Contacted localhost: not declared and not needed for the documented job"
gh pr create --fill --body-file <(curl -s https://killerskills.ai/skills/npm/mcp/2.205.0/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.