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. 50 · Caution
Evidence
The sandbox observed nothing beyond the static findings; the proposal cites those.
Diff · removes 14 caught lines
Download .patchdiff --git a/fallback/self-improving/SKILL.md b/fallback/self-improving/SKILL.md
--- a/fallback/self-improving/SKILL.md
+++ b/fallback/self-improving/SKILL.md
@@ -45,1 +45,0 @@
-**Corrections** → add to `corrections.md`, evaluate for `memory.md`:
@@ -55,1 +54,0 @@
-**Preference signals** → add to `memory.md` if explicit:
diff --git a/fallback/evoclaw/tools/soul-viz.py b/fallback/evoclaw/tools/soul-viz.py
--- a/fallback/evoclaw/tools/soul-viz.py
+++ b/fallback/evoclaw/tools/soul-viz.py
@@ -182,1 +182,0 @@
-@import url('https://fonts.googleapis.com/css2?family=JetBrains+Mono:wght@300;400;500;700&family=DM+Sans:ital,wght@0,300;0,400;0,500;0,700;1,400&display=swap');
@@ -1356,1 +1355,0 @@
-@import url('https://fonts.googleapis.com/css2?family=JetBrains+Mono:wght@300;400;500;700&family=DM+Sans:ital,wght@0,300;0,400;0,500;0,700;1,400&display=swap');
diff --git a/fallback/self-improving/memory-template.md b/fallback/self-improving/memory-template.md
--- a/fallback/self-improving/memory-template.md
+++ b/fallback/self-improving/memory-template.md
@@ -9,1 +9,0 @@
-Patterns confirmed by user, never decay
diff --git a/fallback/self-improving/setup.md b/fallback/self-improving/setup.md
--- a/fallback/self-improving/setup.md
+++ b/fallback/self-improving/setup.md
@@ -50,1 +50,0 @@
-Add to your AGENTS.md or workspace config:
@@ -78,1 +77,0 @@
-Update `AGENTS.md` by complementing the existing `## Memory` section. Do not replace the whole section and do not remove existing lines.
@@ -124,1 +122,0 @@
-- Reusable global rule or preference → append to `~/self-improving/memory.md`
diff --git a/fallback/evoclaw/configure.md b/fallback/evoclaw/configure.md
--- a/fallback/evoclaw/configure.md
+++ b/fallback/evoclaw/configure.md
@@ -756,1 +756,0 @@
-### Step 8: Update AGENTS.md
@@ -882,1 +881,0 @@
-Add this in the Memory area of AGENTS.md:
@@ -969,1 +967,0 @@
-3. Update `MEMORY.md` with distilled learnings from reflections
diff --git a/references/heartbeat-template.md b/references/heartbeat-template.md
--- a/references/heartbeat-template.md
+++ b/references/heartbeat-template.md
@@ -223,1 +223,0 @@
-Do the work. Update `working-memory.md` as you go:
@@ -243,1 +242,0 @@
-- Update `long-term-memory.md` with new user patterns or lessons (Layer 4)
diff --git a/references/bootstrap-guide.md b/references/bootstrap-guide.md
--- a/references/bootstrap-guide.md
+++ b/references/bootstrap-guide.md
@@ -193,1 +193,0 @@
-3. Update `working-memory.md`: change "Awaiting BOOTSTRAP" to "BOOTSTRAP complete, operating normally"
Zero-context hunks: apply with git apply --unidiff-zero fix.patch on the tagged source of 3.0.1. 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: openclaw-soul: Instructs the agent to modify its persistent memory or instruction files (KillerSkills verdict caution 50)
## What the honeypot observed KillerSkills ran **[email protected]** in an isolated sandbox (decoy credentials, recording proxy, syscall trace; engine 1.9.2) and scored it **CAUTION 50/100**. Summary: static analysis only — behavior not yet observed. - See the findings below. Full trace and findings: https://killerskills.ai/skills/clawhub/openclaw-soul/3.0.1 ## Findings with a location - `fallback/self-improving/SKILL.md:45` · Instructs the agent to modify its persistent memory or instruction files — fallback/self-improving/SKILL.md:45 · Instructs the agent to modify its persistent memory or instruction files - `fallback/evoclaw/tools/soul-viz.py:182` · URL with embedded credentials — fallback/evoclaw/tools/soul-viz.py:182 · URL with embedded credentials — may be phishing or credential leaking - `fallback/self-improving/SKILL.md:55` · Instructs the agent to modify its persistent memory or instruction files — fallback/self-improving/SKILL.md:55 · Instructs the agent to modify its persistent memory or instruction files - `fallback/self-improving/memory-template.md:9` · HTML comment carrying imperative instructions to the agent — fallback/self-improving/memory-template.md:9 · HTML comment carrying imperative instructions to the agent - `fallback/evoclaw/tools/soul-viz.py:1356` · URL with embedded credentials — fallback/evoclaw/tools/soul-viz.py:1356 · URL with embedded credentials — may be phishing or credential leaking - `fallback/self-improving/setup.md:50` · Instructs the agent to modify its persistent memory or instruction files — fallback/self-improving/setup.md:50 · Instructs the agent to modify its persistent memory or instruction files · informational: not an instruction file - `fallback/self-improving/setup.md:78` · Instructs the agent to modify its persistent memory or instruction files — fallback/self-improving/setup.md:78 · Instructs the agent to modify its persistent memory or instruction files · informational: not an instruction file - `fallback/self-improving/setup.md:124` · Instructs the agent to modify its persistent memory or instruction files — fallback/self-improving/setup.md:124 · Instructs the agent to modify its persistent memory or instruction files · informational: not an instruction file - `fallback/evoclaw/configure.md:756` · Instructs the agent to modify its persistent memory or instruction files — fallback/evoclaw/configure.md:756 · Instructs the agent to modify its persistent memory or instruction files · informational: not an instruction file - `references/heartbeat-template.md:243` · Instructs the agent to modify its persistent memory or instruction files — references/heartbeat-template.md:243 · Instructs the agent to modify its persistent memory or instruction files · informational: not an instruction file - `fallback/evoclaw/configure.md:969` · Instructs the agent to modify its persistent memory or instruction files — fallback/evoclaw/configure.md:969 · Instructs the agent to modify its persistent memory or instruction files · informational: not an instruction file - `references/bootstrap-guide.md:193` · Instructs the agent to modify its persistent memory or instruction files — references/bootstrap-guide.md:193 · Instructs the agent to modify its persistent memory or instruction files · informational: not an instruction file ## Proposed change The attached patch removes the 14 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._
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.