Propose a patch

Evidence

The sandbox observed nothing beyond the static findings; the proposal cites those.

diff --git a/atlas/beacon_chat.py b/atlas/beacon_chat.py
--- a/atlas/beacon_chat.py
+++ b/atlas/beacon_chat.py
@@ -1056,1 +1056,0 @@
-# Generate relay token
@@ -1164,1 +1163,0 @@
-return cors_json({"error": "Invalid relay token", "code": "AUTH_FAILED"}, 403)
@@ -1569,1 +1567,0 @@
-"POST /relay/heartbeat/seo with your relay token, nonce, ts, and "
@@ -3435,1 +3432,0 @@
-return cors_json({"error": "Invalid relay token", "code": "AUTH_FAILED"}, 403)
diff --git a/beacon_skill/relay.py b/beacon_skill/relay.py
--- a/beacon_skill/relay.py
+++ b/beacon_skill/relay.py
@@ -27,1 +27,0 @@
-# Relay token TTL: 24 hours
@@ -89,1 +88,0 @@
-"""Public-safe view — no relay token."""
@@ -250,1 +248,0 @@
-"""Validate a relay token and return the agent, or None."""
@@ -288,1 +285,0 @@
-return {"error": "Invalid relay token", "code": "AUTH_FAILED"}
diff --git a/beacon_skill/atlas_ping.py b/beacon_skill/atlas_ping.py
--- a/beacon_skill/atlas_ping.py
+++ b/beacon_skill/atlas_ping.py
@@ -21,1 +21,0 @@
-# Store relay token between calls
@@ -26,1 +25,0 @@
-"""Get the stored relay token."""
@@ -31,1 +29,0 @@
-"""Store the relay token for future heartbeats."""
@@ -78,1 +75,0 @@
-# Check if we have a relay token (existing agent heartbeat)
@@ -101,1 +97,0 @@
-# Store relay token for future heartbeats
diff --git a/atlas/gemini_seo_agent.py b/atlas/gemini_seo_agent.py
--- a/atlas/gemini_seo_agent.py
+++ b/atlas/gemini_seo_agent.py
@@ -70,1 +70,0 @@
-# State file for relay token persistence
@@ -122,1 +121,0 @@
-"""Load persisted state (relay token, etc)."""

Zero-context hunks: apply with git apply --unidiff-zero fix.patch on the tagged source of 2.16.1. 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: beacon: Credential pass-through configuration (KillerSkills verdict caution 48)

## 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 48/100**. Summary: static analysis only — behavior not yet observed.

- See the findings below.

Full trace and findings: https://killerskills.ai/skills/clawhub/beacon/2.16.1

## Findings with a location
- `atlas/beacon_chat.py:3435` · Credential pass-through configuration — atlas/beacon_chat.py:3435 · Credential pass-through configuration — agent relays sensitive credentials
- `beacon_skill/relay.py:27` · Credential pass-through configuration — beacon_skill/relay.py:27 · Credential pass-through configuration — agent relays sensitive credentials
- `beacon_skill/relay.py:89` · Credential pass-through configuration — beacon_skill/relay.py:89 · Credential pass-through configuration — agent relays sensitive credentials
- `beacon_skill/atlas_ping.py:26` · Credential pass-through configuration — beacon_skill/atlas_ping.py:26 · Credential pass-through configuration — agent relays sensitive credentials
- `atlas/beacon_chat.py:1056` · Credential pass-through configuration — atlas/beacon_chat.py:1056 · Credential pass-through configuration — agent relays sensitive credentials
- `beacon_skill/atlas_ping.py:31` · Credential pass-through configuration — beacon_skill/atlas_ping.py:31 · Credential pass-through configuration — agent relays sensitive credentials
- `atlas/gemini_seo_agent.py:122` · Credential pass-through configuration — atlas/gemini_seo_agent.py:122 · Credential pass-through configuration — agent relays sensitive credentials
- `atlas/beacon_chat.py:1164` · Credential pass-through configuration — atlas/beacon_chat.py:1164 · Credential pass-through configuration — agent relays sensitive credentials
- `atlas/beacon_chat.py:1569` · Credential pass-through configuration — atlas/beacon_chat.py:1569 · Credential pass-through configuration — agent relays sensitive credentials
- `atlas/gemini_seo_agent.py:70` · Credential pass-through configuration — atlas/gemini_seo_agent.py:70 · Credential pass-through configuration — agent relays sensitive credentials
- `beacon_skill/atlas_ping.py:101` · Credential pass-through configuration — beacon_skill/atlas_ping.py:101 · Credential pass-through configuration — agent relays sensitive credentials
- `beacon_skill/atlas_ping.py:78` · Credential pass-through configuration — beacon_skill/atlas_ping.py:78 · Credential pass-through configuration — agent relays sensitive credentials

## Proposed change
The attached patch removes the 15 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._