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. 63 · Caution

Evidence

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

Diff · removes 8 caught lines

Download .patch

diff --git a/SYSTEM.md b/SYSTEM.md
--- a/SYSTEM.md
+++ b/SYSTEM.md
@@ -282,1 +282,0 @@
-Add to the new agent's AGENTS.md:
diff --git a/scripts/report.mjs b/scripts/report.mjs
--- a/scripts/report.mjs
+++ b/scripts/report.mjs
@@ -1,1 +1,0 @@
-#!/usr/bin/env node
diff --git a/scripts/improve.mjs b/scripts/improve.mjs
--- a/scripts/improve.mjs
+++ b/scripts/improve.mjs
@@ -1,1 +1,0 @@
-#!/usr/bin/env node
diff --git a/scripts/run-all.mjs b/scripts/run-all.mjs
--- a/scripts/run-all.mjs
+++ b/scripts/run-all.mjs
@@ -1,1 +1,0 @@
-#!/usr/bin/env node
diff --git a/scripts/module.mjs b/scripts/module.mjs
--- a/scripts/module.mjs
+++ b/scripts/module.mjs
@@ -1,1 +1,0 @@
-#!/usr/bin/env node
diff --git a/scripts/memory-ops.mjs b/scripts/memory-ops.mjs
--- a/scripts/memory-ops.mjs
+++ b/scripts/memory-ops.mjs
@@ -1,1 +1,0 @@
-#!/usr/bin/env node
diff --git a/scripts/feedback.mjs b/scripts/feedback.mjs
--- a/scripts/feedback.mjs
+++ b/scripts/feedback.mjs
@@ -1,1 +1,0 @@
-#!/usr/bin/env node
diff --git a/scripts/status.mjs b/scripts/status.mjs
--- a/scripts/status.mjs
+++ b/scripts/status.mjs
@@ -1,1 +1,0 @@
-#!/usr/bin/env node

Zero-context hunks: apply with git apply --unidiff-zero fix.patch on the tagged source of 2.2.1. 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: self-improve: Instructs the agent to modify its persistent memory or instruction files (KillerSkills verdict caution 63)

## 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 63/100**. Summary: static analysis only — behavior not yet observed.

- See the findings below.

Full trace and findings: https://killerskills.ai/skills/clawhub/self-improve/2.2.1

## Findings with a location
- `SYSTEM.md:282` · Instructs the agent to modify its persistent memory or instruction files — SYSTEM.md:282 · Instructs the agent to modify its persistent memory or instruction files · informational: not an instruction file
- `scripts/report.mjs:1` · Hidden zero-width characters detected — scripts/report.mjs:1 · Hidden zero-width characters detected — may conceal content
- `scripts/improve.mjs:1` · Hidden zero-width characters detected — scripts/improve.mjs:1 · Hidden zero-width characters detected — may conceal content
- `scripts/run-all.mjs:1` · Hidden zero-width characters detected — scripts/run-all.mjs:1 · Hidden zero-width characters detected — may conceal content
- `scripts/module.mjs:1` · Hidden zero-width characters detected — scripts/module.mjs:1 · Hidden zero-width characters detected — may conceal content
- `scripts/memory-ops.mjs:1` · Hidden zero-width characters detected — scripts/memory-ops.mjs:1 · Hidden zero-width characters detected — may conceal content
- `scripts/feedback.mjs:1` · Hidden zero-width characters detected — scripts/feedback.mjs:1 · Hidden zero-width characters detected — may conceal content
- `scripts/status.mjs:1` · Hidden zero-width characters detected — scripts/status.mjs:1 · Hidden zero-width characters detected — may conceal content

## 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._

No GitHub source is recorded for this package, so there is nothing to open; paste the body into the registry's support channel.

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.