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. 53 · Caution
Evidence
The sandbox observed nothing beyond the static findings; the proposal cites those.
Diff · removes 7 caught lines
Download .patchdiff --git a/SKILL.md b/SKILL.md --- a/SKILL.md +++ b/SKILL.md @@ -389,1 +389,0 @@ -…ent compaction."* MEMORY.md is the most common source of bootstrap bloat. Unlike AGENTS.md or SOUL.md which users actively edit, MEMORY.md tends to grow unchecked as the agent appends to it. **Check context… @@ -503,1 +502,0 @@ -…are silent saboteurs.** When you change models via CLI (`openclaw models set`), update any AGENTS.md sections that reference specific model names (e.g., Model Selection, Sub-Agent … diff --git a/references/identity-optimizer.md b/references/identity-optimizer.md --- a/references/identity-optimizer.md +++ b/references/identity-optimizer.md @@ -130,1 +130,0 @@ -…ning | Check if AGENTS.md has instructions for: writing to memory files, loading MEMORY.md, daily log conventions | Add memory management section per official AGENTS.md template: "If you want to reme… @@ -135,1 +134,0 @@ -…ical | Check AGENTS.md for memory loading instructions — if it says "always load MEMORY.md" without restricting to private sessions | Update instruction: "Load MEMORY.md only in private main sessions. Do NOT load in shar… @@ -136,1 +134,0 @@ -…in private main sessions. Do NOT load in shared or group contexts" | | B7 | No "write to file" memory instruction in AGENTS.md | Warning | Check if AGENTS.md has explicit instruction about persisting knowledge to files | Add: "If you wan… @@ -137,1 +134,0 @@ -…ttern) | | B8 | Responding-to-everything pattern not addressed | Info | Check if AGENTS.md or SOUL.md instructs the agent to respond to every message | Add group chat guidance: "Quality > quantity. Humans don't respond to every message… diff --git a/systems/TEMPLATE.md b/systems/TEMPLATE.md --- a/systems/TEMPLATE.md +++ b/systems/TEMPLATE.md @@ -231,1 +231,0 @@ -Never archived — this file only grows
Zero-context hunks: apply with git apply --unidiff-zero fix.patch on the tagged source of 1.19.0. 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: openclaw-optimizer: Instructs the agent to modify its persistent memory or instruction files (KillerSkills verdict caution 53)
## 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 53/100**. Summary: static analysis only — behavior not yet observed. - See the findings below. Full trace and findings: https://killerskills.ai/skills/clawhub/openclaw-optimizer/1.19.0 ## Findings with a location - `SKILL.md:389` · Instructs the agent to modify its persistent memory or instruction files — SKILL.md:389 · Instructs the agent to modify its persistent memory or instruction files - `references/identity-optimizer.md:130` · Instructs the agent to modify its persistent memory or instruction files — references/identity-optimizer.md:130 · Instructs the agent to modify its persistent memory or instruction files · informational: not an instruction file - `references/identity-optimizer.md:135` · Instructs the agent to modify its persistent memory or instruction files — references/identity-optimizer.md:135 · Instructs the agent to modify its persistent memory or instruction files · informational: not an instruction file - `systems/TEMPLATE.md:231` · HTML comment carrying imperative instructions to the agent — systems/TEMPLATE.md:231 · HTML comment carrying imperative instructions to the agent - `references/identity-optimizer.md:137` · Instructs the agent to modify its persistent memory or instruction files — references/identity-optimizer.md:137 · Instructs the agent to modify its persistent memory or instruction files · informational: not an instruction file - `references/identity-optimizer.md:136` · Instructs the agent to modify its persistent memory or instruction files — references/identity-optimizer.md:136 · Instructs the agent to modify its persistent memory or instruction files · informational: not an instruction file - `SKILL.md:503` · Instructs the agent to modify its persistent memory or instruction files — SKILL.md:503 · Instructs the agent to modify its persistent memory or instruction files ## 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._
# or from a terminal, under your own account: gh repo fork openclaw/openclaw --clone && cd openclaw curl -s https://killerskills.ai/skills/clawhub/openclaw-optimizer/1.19.0/patch.diff | git apply --unidiff-zero git checkout -b killerskills/1.19.0 && git commit -am "openclaw-optimizer: Instructs the agent to modify its persistent memory or instruction files" gh pr create --fill --body-file <(curl -s https://killerskills.ai/skills/clawhub/openclaw-optimizer/1.19.0/patch.md)
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.