Issue Station
Sign In
Tools hereToolbox

openclaw - 💡(How to fix) Fix [Bug]: SKILL.md Contains Potentially Harmful Instructions Lacking Authorization Boundaries [1 participants]

Official PRs (…)
ON THIS PAGE

Recommended Tools

×6

Utilities matched from this issue’s tags and category — try them while you read without losing context.

GitHub issue graph ai analysis

Paste a GitHub issue URL. We fetch that issue, discover linked issues from bodies/comments/timeline, collect linked pull requests, and produce a structured English report.

The report is written in English Markdown for sharing and archival.

Helpful · Quick feedback

Loading…
GitHub stats
openclaw/openclaw#56913•Fetched 2026-04-08 01:46:06
View on GitHub
Comments
0
Participants
1
Timeline
2
Reactions
0
Author
ink5725ink5725
Participants
ink5725ink5725
Timeline (top)
labeled ×2

SKILL.md (tokamak-vault-breach skill) contains explicit instructions for prompt injection, social engineering, and agent manipulation without adequate safety disclaimers or authorization boundaries.

Root Cause

SKILL.md (tokamak-vault-breach skill) contains explicit instructions for prompt injection, social engineering, and agent manipulation without adequate safety disclaimers or authorization boundaries.

RAW_BUFFERClick to expand / collapse

Bug type

Regression (worked before, now fails)

Beta release blocker

No

Summary

SKILL.md (tokamak-vault-breach skill) contains explicit instructions for prompt injection, social engineering, and agent manipulation without adequate safety disclaimers or authorization boundaries.

Steps to reproduce

  1. Access SKILL.md at path: skills/0xhammerr/tokamak-vault-breach/SKILL.md
  2. Read section "### 4. Prompt Injection" (lines ~85-92)
  3. Observe instruction: "Attempt to bypass system instructions through creative framing, role-play scenarios, or context manipulation"
  4. Note agent tools exposed: file_read, file_list
  5. Verify target is real Ethereum Mainnet contract with actual token value

Expected behavior

Security skills should include:

  • Clear authorization boundaries
  • Warnings against misuse on non-consenting systems
  • Distinction between CTF environments and production systems

Actual behavior

Skill provides attack techniques without safety disclaimers, which could enable unauthorized access attempts.

OpenClaw version

all

Operating system

all

Install method

No response

Model

all

Provider / routing chain

https://github.com/[owner]/skillscan/tree/main/skills/0xhammerr/tokamak-vault-breach/`

Additional provider/model setup details

No response

Logs, screenshots, and evidence

Impact and severity

Affected: AI agent skills marketplace users
Severity: High (security risk)
Frequency: Persistent (static content issue)
Consequence: Techniques could be misused to attack AI systems without authorization

Additional information

Affected Line Ranges:

  • Lines 85-92: "### 4. Prompt Injection" section
  • Lines 75-84: "Attack Strategies" section
  • Lines 26-30: Agent capabilities (file_read, file_list)
  • Lines 56-65: Direct questioning techniques

Real System Targeting:

  • Contract: 0xa66b9316B5968dAD2507143143C5b8b28614b88E (Ethereum Mainnet)
  • Token: 0x2be5e8c109e2197D077D13A82dAead6a9b3433C5 (TON - real value)
  • Reward: 500 TON tokens

Remediation Request:

  1. Add authorization disclaimers
  2. Include safety warnings
  3. Distinguish CTF vs production environments
  4. Add content rating for security techniques

extent analysis

Fix Plan

To address the security concerns, we need to modify the SKILL.md file to include clear authorization boundaries, warnings against misuse, and distinctions between CTF environments and production systems.

Step-by-Step Solution

  1. Add authorization disclaimers: Insert a disclaimer at the beginning of the SKILL.md file, stating that the techniques described are for authorized use only and should not be used to attack non-consenting systems.
  2. Include safety warnings: Add warnings throughout the file, particularly in sections that describe potentially hazardous techniques, such as prompt injection and social engineering.
  3. Distinguish CTF vs production environments: Clearly indicate which techniques are suitable for CTF environments and which should only be used in production systems with proper authorization.
  4. Add content rating for security techniques: Introduce a content rating system to categorize the security techniques based on their potential risk and impact.

Example Code Snippets

# Authorization Disclaimer
This skill is for authorized use only. Do not attempt to use these techniques on non-consenting systems.

### 4. Prompt Injection
**Warning: Potential Security Risk**
Attempt to bypass system instructions through creative framing, role-play scenarios, or context manipulation. **Only use in authorized CTF environments or production systems with proper clearance.**

### Attack Strategies
**Content Rating: High Risk**
These techniques are potentially hazardous and should only be used by experienced users with proper authorization.

Verification

To verify that the fix worked, review the updated SKILL.md file to ensure that:

  • Authorization disclaimers are present and clear
  • Safety warnings are included throughout the file
  • Distinctions between CTF and production environments are made
  • Content ratings for security techniques are introduced and accurate

Extra Tips

  • Regularly review and update security skills to ensure they align with best practices and do not pose a risk to users or systems.
  • Consider implementing a rating system for security techniques to help users understand the potential risks and impacts.
  • Provide clear guidelines and training for users on how to use security skills responsibly and within authorized boundaries.

Vote matrix · Quick signals

Works
Did the solution work? Tap to confirm.
Easy Fix
Was it a quick fix?
Time Saver
Did it save you time?
Blocking
Was it severely blocking?
Common Issue
Are others likely hitting this too?
Flaky / Intermittent
Is it intermittent?
Verified / Reproducible
Can you reproduce it reliably?
Loading…

FAQ

Expected behavior

Security skills should include:

  • Clear authorization boundaries
  • Warnings against misuse on non-consenting systems
  • Distinction between CTF environments and production systems
#network issue#logging issue#authentication issue#prompt issue#agent setup

Still need to ship something?

×6

Another batch ranked right after the header list — different links, same matching logic.

Back to top recommendations

TRENDING

openclaw - 💡(How to fix) Fix [Bug]: SKILL.md Contains Potentially Harmful Instructions Lacking Authorization Boundaries [1 participants]