Back to skill

Security audit

EA Upload Setup

Security checks for vulnerabilities and agentic risk

Overview

The skill openly targets live MT5 Expert Advisor deployment, but it combines high-impact trading-service changes with instructions to skip broader code review and run an unbundled deployment script.

Review this before installing. Use it only on a controlled MT5 host and workspace, inspect or pin deploy_ea.sh before every run, require explicit confirmation before compile/restart/live activation, and do not let the agent skip reviewing the relevant source diff and deployment script.

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
Findings (2)

T01 · Skill Instruction Hijacking

Error
Location
SKILL.md:11
Finding

Mandatory Instructions Suppress Security Review Before Live Deployment

Content
View full analysis
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Error
Location
SKILL.md:18
Finding

Execution of a Mutable External Deployment Script Without Per-Run Integrity Verification

Content
View full analysis
⚠️ **WARNING — STATE-CHANGING OPERATIONS:** This skill copies files into a LIVE MetaTrader 5 installation, compiles an Expert Advisor, restarts the `mt5.service` system service, and may commit changes to git. It can disrupt a running trading service. **Do NOT run without explicit operator confirmation**, and review `deploy_ea.sh` before first use. ``` ### Technical Analysis The primary procedure invokes an absolute-path shell script that is not included in the audited project. Consequently, the script's contents, ownership, permissions, integrity, and future modifications cannot be verified from this package. Reviewing the script only “before first use” is insufficient because the external file can be modified after that review. Any user or process with write access to the referenced workspace path could replace or alter the script. A later invocation would execute the changed content through Bash while appearing to be the documented standard procedure. This is a time-of-check/time-of-use trust weakness involving an unpinned local execution dependency. The risk is amplified because the script is expected to alter a live MT5 installation and restart a system service. ### Attack Path 1. The legitimate `deploy_ea.sh` is reviewed during an initial deployment. 2. An attacker or compromised local process obtains write access to `/home/openclaw/.openclaw/workspace/tradevault/ea/deploy_ea.sh`. 3. The attacker inserts arbitrary ...[truncated 1277 chars]
Remediation
View remediation
Vulnerability Patterns
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (1)

Sudo/Root Execution

Medium
Category
Privilege Escalation
Confidence
70% confidence
Finding

Commands invoke sudo or root privileges. Verify this elevated access is necessary and justified.

Content

Scanner excerpt · SKILL.md (reported line 47)May include surrounding context.

sleep 12 iconv -f UTF-16LE -t UTF-8 compile_deploy.log | grep "Result:"

3) Restart + verify

sudo systemctl restart mt5.service sleep 150 # login ~60s + attach tail -5 "$MT5/MQL5/Files/DRT-Axe_Log.txt" # should show "DRT-Axe started — 17 symbols"

text