Security audit

anchored-vwap-scalper

Security checks for vulnerabilities and agentic risk

Overview

This is a real automated crypto-trading skill with live order capability, but its safety claims and controls are not clear or reliable enough for live funds without review.

Treat this as a Review item before installation. Use dry-run only unless you have independently tested the strategy, verified exchange order semantics, restricted the API key to the minimum trading permissions, removed withdrawal authority if possible, and are willing to risk the configured account collateral. Do not rely on the advertised 1% risk, daily loss, leverage, stop-loss, or take-profit claims without code fixes and live-exchange validation.

A.I.G

Vulnerability Patterns

Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
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 (3)

T09 · Insecure Skill Coding Practices

Error

Location: scripts/strategy_engine.py:116
Finding: Protective Order Failures Can Leave Live Leveraged Positions Unprotected
Content: View full analysis
Remediation: View remediation

T09 · Insecure Skill Coding Practices

Error

Location: scripts/strategy_engine.py:30
Finding: Declared Daily-Loss and Leverage Controls Are Not Enforced
Content: View full analysis
0: logging.info(f"🚀 BTC SIGNAL: {signal['strategy']} | {side.upper()} | AVWAP {avwap:.2f} | Price {current_price:.2f}") entry_order = self.place_order(side=side, order_type='market', amount=contracts) ``` The state saved after execution only updates the bar timestamp: ```python self.state["last_bar_ts"] = str(df['timestamp'].iloc[-1]) self._save_state() return signals ``` ### Technical Analysis The configuration declares a 3% daily-loss limit and leverage of 10, and the default state contains a `daily_pnl` field. However, the reviewed execution flow does not: - Retrieve authoritative realized and unrealized daily profit and loss. - Update `daily_pnl`. - Compare current losses against `daily_loss_limit`. - Reject new entries after the loss limit is reached. - Reset and reconcile the daily metric by trading day. - Set leverage through the exchange API. - Verify that ac ...[truncated 1720 chars]
Remediation: View remediation

T09 · Insecure Skill Coding Practices

Warning

Location: scripts/strategy_engine.py:99
Finding: Minimum-Order Clamping Can Exceed the Intended Per-Trade Risk
Content: View full analysis
Remediation: View remediation

SkillSpector

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands

Findings (8)

Tp4

High

Category: MCP Tool Poisoning
Confidence: 98% confidence
Finding: This second mismatch finding likewise shows the skill claims anchored VWAP, four strategies, and explicit 1% risk controls, while the underlying behavior reportedly only loads credentials and initializes an exchange client. That is dangerous because users may enable the skill with live API keys under the false assumption that the advertised trading logic and safeguards are present, leading to unsafe or unintended operation.
Content: No source excerpt is available for this finding.

Tp4

High

Category: MCP Tool Poisoning
Confidence: 98% confidence
Finding: This second mismatch finding likewise shows the skill claims anchored VWAP, four strategies, and explicit 1% risk controls, while the underlying behavior reportedly only loads credentials and initializes an exchange client. That is dangerous because users may enable the skill with live API keys under the false assumption that the advertised trading logic and safeguards are present, leading to unsafe or unintended operation.
Content: No source excerpt is available for this finding.

Undeclared Tool Scope

Medium

Category: MCP Least Privilege
Confidence: 86% confidence
Finding: The skill declares required environment variables and appears to rely on file-writing/state persistence, but it does not declare an explicit tool scope such as permissions or allowed-tools. That creates a trust and review gap: operators may approve the skill without understanding it can access secrets or write local state, which is especially risky in a trading context handling private keys.
Content: No source excerpt is available for this finding.

Intent-Code Divergence

Medium

Category: Not specified by scanner
Confidence: 96% confidence
Finding: The documentation is internally inconsistent about whether both fixed stop-loss and take-profit are provided or only fixed stop-loss. In an automated trading skill, inconsistent safety-control documentation can directly mislead operators about exit protections and materially change financial risk assumptions.
Content: No source excerpt is available for this finding.

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 94% confidence
Finding: This code reads a private API key and account identifiers from environment variables, which is a sensitive operation covered by the missing-warning rule for code files. The file contains no confirmation prompt, logging, comment, or docstring disclosing that credentials are accessed, and the stated purpose is not documented in this file.
Content: No source excerpt is available for this finding.

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 89% confidence
Finding: The code can place live market, stop, and take-profit orders when DRY_RUN is false, but it contains no in-code warning, safety acknowledgement, or confirmation barrier around that behavior. In a trading skill, this is dangerous because a user or integrator may enable live mode via environment variables and trigger real leveraged trades on a production exchange without realizing the file is capable of executing them directly.
Content: No source excerpt is available for this finding.

Intent-Code Divergence

Low

Category: Not specified by scanner
Confidence: 96% confidence
Finding: The documentation gives conflicting risk-control behavior: the metadata describes 'fixed SL + TP' while the body later states 'fixed SL only'. In an automated trading skill, inconsistent stop-loss/take-profit semantics can cause an operator to deploy the strategy under false assumptions about exit controls, increasing the chance of unmanaged positions or unexpected loss behavior.
Content: No source excerpt is available for this finding.

Missing User Warnings

Low

Category: Not specified by scanner
Confidence: 88% confidence
Finding: This code persists state to ../state/avwap_state.json, which is a file write affecting the local filesystem. There is no confirmation prompt, logging/print statement, or explanatory docstring/comment disclosing that the skill creates and updates a state file.
Content: No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.