Propose a patch

Evidence

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

diff --git a/src/layered_memory_mcp/config.py b/src/layered_memory_mcp/config.py
--- a/src/layered_memory_mcp/config.py
+++ b/src/layered_memory_mcp/config.py
@@ -484,1 +484,0 @@
-# v3.3.0 new field — MEMORY.md write-guard install policy
@@ -746,1 +745,0 @@
-# v3.2.0: Agent-memory write mode — how dual-write treats MEMORY.md:
@@ -763,1 +761,0 @@
-# v3.3.0: MEMORY.md write guard — install policy for the Hermes
diff --git a/src/layered_memory_mcp/server.py b/src/layered_memory_mcp/server.py
--- a/src/layered_memory_mcp/server.py
+++ b/src/layered_memory_mcp/server.py
@@ -2539,1 +2539,0 @@
-# v3.3.0: MEMORY.md write guard. Always detected; installed only under
@@ -2636,1 +2635,0 @@
-- "guard_status": report the MEMORY.md write-guard status
@@ -2637,1 +2635,0 @@
-- "install_guard": deploy + enable the MEMORY.md write-guard plugin and
diff --git a/src/layered_memory_mcp/write_guard/plugin/__init__.py b/src/layered_memory_mcp/write_guard/plugin/__init__.py
--- a/src/layered_memory_mcp/write_guard/plugin/__init__.py
+++ b/src/layered_memory_mcp/write_guard/plugin/__init__.py
@@ -1,1 +1,0 @@
-"""Hermes plugin entry point for the layered-memory MEMORY.md write guard.

Zero-context hunks: apply with git apply --unidiff-zero fix.patch on the tagged source of 3.5.1. 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: layered-memory-mcp: Instructs the agent to modify its persistent memory or instruction files (KillerSkills verdict caution 79)

## What the honeypot observed
KillerSkills ran **[email protected]** in an isolated sandbox (decoy credentials, recording proxy, syscall trace; engine 1.6) and scored it **CAUTION 79/100**. Summary: no network, touches nothing outside the project, writes no memory.

- See the findings below.

Full trace and findings: https://killerskills.ai/skills/pypi/layered-memory-mcp/3.5.1

## Findings with a location
- `src/layered_memory_mcp/config.py:746` · Instructs the agent to modify its persistent memory or instruction files — src/layered_memory_mcp/config.py:746 · Instructs the agent to modify its persistent memory or instruction files · informational: not an instruction file
- `src/layered_memory_mcp/config.py:763` · Instructs the agent to modify its persistent memory or instruction files — src/layered_memory_mcp/config.py:763 · Instructs the agent to modify its persistent memory or instruction files · informational: not an instruction file
- `src/layered_memory_mcp/config.py:484` · Instructs the agent to modify its persistent memory or instruction files — src/layered_memory_mcp/config.py:484 · Instructs the agent to modify its persistent memory or instruction files · informational: not an instruction file
- `src/layered_memory_mcp/server.py:2539` · Instructs the agent to modify its persistent memory or instruction files — src/layered_memory_mcp/server.py:2539 · Instructs the agent to modify its persistent memory or instruction files · informational: not an instruction file
- `src/layered_memory_mcp/server.py:2636` · Instructs the agent to modify its persistent memory or instruction files — src/layered_memory_mcp/server.py:2636 · Instructs the agent to modify its persistent memory or instruction files · informational: not an instruction file
- `src/layered_memory_mcp/write_guard/plugin/__init__.py:1` · Instructs the agent to modify its persistent memory or instruction files — src/layered_memory_mcp/write_guard/plugin/__init__.py:1 · Instructs the agent to modify its persistent memory or instruction files · informational: not an instruction file
- `src/layered_memory_mcp/server.py:2637` · Instructs the agent to modify its persistent memory or instruction files — src/layered_memory_mcp/server.py:2637 · Instructs the agent to modify its persistent memory or instruction files · informational: not an instruction file

## Proposed change
The attached patch removes the 7 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._