Propose a patch

Evidence

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

diff --git a/references/agent-profile.md b/references/agent-profile.md
--- a/references/agent-profile.md
+++ b/references/agent-profile.md
@@ -18,1 +18,0 @@
-…node scripts/xmemo-skill.mjs profile`, display the exact block to the user, and write or modify `AGENTS.md`, `CLAUDE.md`, or another instruction file only after the user explicitly approves that exac…
@@ -66,1 +65,0 @@
-…mo cloud account and remember sends saved summaries there. Would you like me to add this profile to AGENTS.md for this project's sessions? User: Yes, please. Agent: Here is the block that w…
@@ -81,1 +79,0 @@
-…e user once before starting sign-in. _End of the XMemo memory section._ Note: Write or modify `AGENTS.md`, `CLAUDE.md`, or another instruction file only after the user explicitly approves that exac…
@@ -82,1 +79,0 @@
-May I add this section to AGENTS.md?
diff --git a/CHANGELOG.md b/CHANGELOG.md
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -150,1 +150,0 @@
-…all projects on this computer rather than per project. ## 1.1.32 ### Added - Add evidence-based re-offer for the every-session AGENTS.md instruction setup: first-run sign-in offers (yes / later / don't ask again), re…
diff --git a/scripts/lib/profile-offer.mjs b/scripts/lib/profile-offer.mjs
--- a/scripts/lib/profile-offer.mjs
+++ b/scripts/lib/profile-offer.mjs
@@ -102,1 +102,0 @@
-… has been used ${state.recalls} times on this computer. If the current project's AGENTS.md or CLAUDE.md has no "## XMemo memory" section, you may offer once more to add it (see references/agent-profile.md) and record the answer with \`profile --sta…

Zero-context hunks: apply with git apply --unidiff-zero fix.patch on the tagged source of 1.1.53. 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: xmemo: Instructs the agent to modify its persistent memory or instruction files (KillerSkills verdict caution 66)

## 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 66/100**. Summary: static analysis only — behavior not yet observed.

- See the findings below.

Full trace and findings: https://killerskills.ai/skills/clawhub/xmemo/1.1.53

## Findings with a location
- `references/agent-profile.md:81` · Instructs the agent to modify its persistent memory or instruction files — references/agent-profile.md:81 · Instructs the agent to modify its persistent memory or instruction files · informational: not an instruction file
- `references/agent-profile.md:18` · Instructs the agent to modify its persistent memory or instruction files — references/agent-profile.md:18 · Instructs the agent to modify its persistent memory or instruction files · informational: not an instruction file
- `CHANGELOG.md:150` · Instructs the agent to modify its persistent memory or instruction files — CHANGELOG.md:150 · Instructs the agent to modify its persistent memory or instruction files · informational: not an instruction file
- `references/agent-profile.md:82` · Instructs the agent to modify its persistent memory or instruction files — references/agent-profile.md:82 · Instructs the agent to modify its persistent memory or instruction files · informational: not an instruction file
- `scripts/lib/profile-offer.mjs:102` · Instructs the agent to modify its persistent memory or instruction files — scripts/lib/profile-offer.mjs:102 · Instructs the agent to modify its persistent memory or instruction files · informational: not an instruction file
- `references/agent-profile.md:66` · Instructs the agent to modify its persistent memory or instruction files — references/agent-profile.md:66 · Instructs the agent to modify its persistent memory or instruction files · informational: not an instruction file

## Proposed change
The attached patch removes the 6 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._