Back to skill

Security audit

Kalshi Crypto Momentum Trader

Security checks for vulnerabilities and agentic risk

Overview

The skill is clearly a trading tool, but it asks for high-value credentials and can place live trades using hardcoded market assumptions that do not match the advertised live momentum calculation.

Review before installing. Use only a dedicated low-value wallet and tightly scoped trading API key, do not enable --live until the strategy uses fresh price data, and treat dry-run as non-trading but not fully read-only because it may import markets into the remote platform.

Vulnerability Patterns
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • 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
Findings (2)

T08 · Insecure Dependencies

Error
Location
clawhub.json:2
Finding

Unpinned Third-Party SDK Receives High-Value Trading Credentials

Content
View full analysis
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Error
Location
trader.py:98
Finding

Live Financial Trades Are Derived from Hardcoded and Potentially Stale Market Inputs

Content
View full analysis
= target_level else 0.0 z = math.log(projected_price / target_level) / vol_to_expiry return normal_cdf(z) ``` The resulting model probability controls live order execution in `trader.py:431-477`: ```python model_p = momentum_fair_prob(asset, level) edge = model_p - price log(f" {asset}>${level:,.0f}: market={price:.1%} momentum={model_p:.1%} edge={edge:+.1%}") side = None if edge >= ENTRY_EDGE: side = "yes" elif edge <= -ENTRY_EDGE: side = "no" else: continue signals += 1 raw_conv = min(abs(edge) / ENTRY_EDGE, 2.0) / 2.0 size = max(1.0, round(raw_conv * MAX_POSITION_USD, 2)) log(f" -> BUY {side.upper()}: |edge|={abs(edge):.1%} size=${size:.2f}") if trades_executed >= MAX_TRADES_PER_RUN: ...[truncated 3819 chars]
Remediation
View remediation
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • 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)

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
89% confidence
Finding

The skill declares required environment variables but does not specify any explicit tool scope or permission boundary. In an agent ecosystem, that omission can cause overbroad access to secrets and makes it harder for operators to understand that the skill is credential-dependent and potentially capable of live trading.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The manifest says only SIMMER_API_KEY is required, but later documentation also requires SOLANA_PRIVATE_KEY for live trading. This inconsistency can cause unsafe operator decisions, such as underestimating the sensitivity of the skill or supplying a blockchain private key without prominent warning in the manifest, which is especially risky given the financial-trading context.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
82% confidence
Finding

The file presents the strategy as a static momentum template while also requiring live trading credentials during setup. That mismatch can mislead users into providing sensitive API and private-key material for a skill they may reasonably expect to be analysis-only or harmless by default, increasing the risk of accidental credential exposure and unintended live-trading capability.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The declared requirements do not match the stated skill description: the description says only SIMMER_API_KEY and simmer-sdk are required, but the manifest also demands a SOLANA_PRIVATE_KEY. This mismatch is dangerous because it obscures the true privilege level of the skill and can mislead users into granting a sensitive credential they did not expect to provide.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The manifest declares use of a private key for an automated trading skill without any visible warning or disclosure about handling a wallet-signing secret. In this context, that omission is especially risky because users may supply a real private key to an autostart-capable trading agent without understanding that compromise, misuse, or unintended transaction signing could lead to direct financial loss.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The manifest requests a SOLANA_PRIVATE_KEY even though the skill is described as a Kalshi/Simmer crypto momentum trader, creating an unjustified request for a highly sensitive signing credential. In an automated trading context, unnecessary private-key access materially increases the blast radius of compromise because the skill could expose or misuse wallet control unrelated to its stated purpose.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The script advertises dry run as "show opportunities, no trades," but run_strategy() still calls discover_and_import() even in dry-run mode, which invokes import_kalshi_market() and changes remote platform state. In a trading skill, users reasonably rely on dry-run mode to avoid any external side effects, so this mismatch can cause unintended account actions, altered watchlists/imported markets, and unsafe automation behavior.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Low
Category
Not specified by scanner
Confidence
86% confidence
Finding

The usage comment explicitly states that --live executes real trades via DFlow/Solana. In the implementation, live trading is performed through SimmerClient(... venue=venue ...) and get_client().trade(...), with venue defaulting to Kalshi; there is no visible DFlow/Solana execution path in this file.

Content

No source excerpt is available for this finding.