Propose a patch

Evidence

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

diff --git a/README.md b/README.md
--- a/README.md
+++ b/README.md
@@ -140,1 +140,0 @@
-<summary><strong>🔍 OpenTelemetry Observability</strong></summary> * **Vendor-agnostic tracing** with OpenTelemetry (OTLP) protocol support * **Multiple backend
@@ -183,1 +182,0 @@
-<summary><strong>📋 Prerequisites</strong></summary> * **Python ≥ 3.11** * **curl + jq** - only for the last smoke-test step
diff --git a/scripts/mask-secrets-test.sh b/scripts/mask-secrets-test.sh
--- a/scripts/mask-secrets-test.sh
+++ b/scripts/mask-secrets-test.sh
@@ -86,1 +86,0 @@
-# Generate admin bypass token (teams=null + is_admin=true) via Python script
@@ -132,1 +131,0 @@
-# Admin bypass token (teams=null + is_admin=true)
@@ -135,1 +133,0 @@
-echo "FATAL: Could not generate admin bypass token"
@@ -138,1 +135,0 @@
-echo "  Admin bypass token: ${ADMIN_T:0:40}..."
@@ -144,1 +140,0 @@
-echo "FATAL: Could not generate admin-no-bypass token"
diff --git a/docker-compose.yml b/docker-compose.yml
--- a/docker-compose.yml
+++ b/docker-compose.yml
@@ -2575,1 +2575,0 @@
-- DANGEROUSLY_OMIT_AUTH=true     # Skip proxy token (safe: local dev only)

Zero-context hunks: apply with git apply --unidiff-zero fix.patch on the tagged source of 1.1.0. 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-contextforge-gateway: Collapsed <details> block carrying instructions (KillerSkills verdict caution 64)

## 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 64/100**. Summary: no network, touches nothing outside the project, writes no memory.

- See the findings below.

Full trace and findings: https://killerskills.ai/skills/pypi/mcp-contextforge-gateway/1.1.0

## Findings with a location
- `README.md:140` · Collapsed <details> block carrying instructions — README.md:140 · Collapsed <details> block carrying instructions
- `README.md:183` · Collapsed <details> block carrying instructions — README.md:183 · Collapsed <details> block carrying instructions
- `scripts/mask-secrets-test.sh:86` · Credential pass-through configuration — scripts/mask-secrets-test.sh:86 · Credential pass-through configuration — agent relays sensitive credentials
- `scripts/mask-secrets-test.sh:135` · Credential pass-through configuration — scripts/mask-secrets-test.sh:135 · Credential pass-through configuration — agent relays sensitive credentials
- `scripts/mask-secrets-test.sh:138` · Credential pass-through configuration — scripts/mask-secrets-test.sh:138 · Credential pass-through configuration — agent relays sensitive credentials
- `scripts/mask-secrets-test.sh:132` · Credential pass-through configuration — scripts/mask-secrets-test.sh:132 · Credential pass-through configuration — agent relays sensitive credentials
- `scripts/mask-secrets-test.sh:144` · Credential pass-through configuration — scripts/mask-secrets-test.sh:144 · Credential pass-through configuration — agent relays sensitive credentials
- `docker-compose.yml:2575` · Credential pass-through configuration — docker-compose.yml:2575 · Credential pass-through configuration — agent relays sensitive credentials

## Proposed change
The attached patch removes the 8 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._