Back to skill

Security audit

Kalshi Crypto Correlation Trader

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed trading bot, but it requests high-value credentials broadly and has dry-run and safety-control behavior that users should review before installing.

Review carefully before installing. Use only limited-scope trading credentials if available, avoid exposing a full wallet private key unless you understand why it is needed, pin and review simmer-sdk before live use, and treat dry-run as not fully read-only because it can modify remote Simmer market state.

Vulnerability Patterns
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
  • 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
Findings (4)

T05 · Unauthorized Access and Privilege Escalation

Error
Location
clawhub.json:3
Finding

Unnecessary Provisioning of the Solana Private Key

Content
View full analysis
SDK tab") sys.exit(1) venue = os.environ.get("TRADING_VENUE", "kalshi") _client = SimmerClient(api_key=api_key, venue=venue, live=live) ``` Consequently, the raw wallet key is requested even for dry-run, configuration-display, and position-display operations that do not require local wallet signing. Any module loaded into the process can inspect environment variables, including the third-party SDK and optional journal modules. The authenticated API key is consistent with remote market and trading operations. No direct credential-exfiltration endpoint or direct private-key read was found in the first-party code. The vulnerability is the unnecessary expansion of the process credential boundary. ### Attack Path 1. A user installs and runs the Skill according to its metadata. 2. The ...[truncated 955 chars]
Remediation
View remediation

T08 · Insecure Dependencies

Error
Location
clawhub.json:8
Finding

Unpinned Trading SDK Executes with High-Value Credentials

Content
View full analysis
SDK tab") sys.exit(1) venue = os.environ.get("TRADING_VENUE", "kalshi") _client = SimmerClient(api_key=api_key, venue=venue, live=live) ``` ### Technical Analysis The package declares `simmer-sdk` without an exact version, lockfile, or integrity hash. Installation can therefore resolve to a future package release whose contents differ from those reviewed by the Skill author. This dependency is security-sensitive because: - `simmer_sdk.skill` executes during module initialization. - `SimmerClient` receives the high-value `SIMMER_API_KEY`. - The process metadata also requests `SOLANA_PRIVATE_KEY`. - The SDK controls authenticated network requests and trade submission. No evidence was found that the currently intended package is itself malicious. The weakness is that the effective code can change after this Skill is audited, while retaining access to trading authority and process credentials. ### Attack Path 1. An attacker compromises the package publisher account, distribution pipeline, or a future dependency release. 2. A user installs the Skill without a version lock. 3. The package installer resolves the attacker-controlled or compromised release. 4. The dependency executes at i ...[truncated 692 chars]
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Error
Location
trader.py:285
Finding

Risk Controls Fail Open When Market Context Retrieval Fails

Content
View full analysis
dict: """Get market context with safeguards and edge analysis.""" try: if my_probability is not None: return get_client()._request( "GET", f"/api/sdk/context/{market_id}", params={"my_probability": my_probability} ) return get_client().get_market_context(market_id) except Exception: return None def check_safeguards(context: dict) -> tuple[bool, list[str]]: """Check context for safeguards. Returns (should_trade, reasons).""" if not context: return True, [] ``` The returned authorization decision is used in the live trade path: ```python ctx = get_market_context( market_id, my_probability=fair_p if side == "yes" else 1 - fair_p ) ok, reasons = check_safeguards(ctx) if not ok: log(f" Skipped: {'; '.join(reasons)}") skip_reasons.extend(reasons) continue ``` ### Technical Analysis `get_market_context` catches every exception and converts the failure into `None`. `check_safeguards` then interprets missing or empty context as approval to trade. This is a fail-open design. Context is the source of checks for: - Resolved markets - Severe flip-flop warnings - Time remaining until resolution - Minimum liquidity - Estimated slippage If the context endpoint is unavailable, returns malformed data, encounters an authentication problem, or raises an SDK exception, all of those validations are bypassed. Both entry and exit execution paths rely on this function. ### Attack Path 1. The Skill is launched with `--live`. 2. A trade signal or exit condition is detected. 3. The context request fails because of a network disruption, API failure, malformed response, SDK error, or deliberate ...[truncated 820 chars]
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Warning
Location
trader.py:463
Finding

Dry-Run Mode Performs Authenticated Remote State Mutation

Content
View full analysis
Remediation
View remediation
Vulnerability Patterns
  • 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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (6)

Tp4

High
Category
MCP Tool Poisoning
Confidence
97% confidence
Finding

The skill description understates operational behavior: beyond signal generation, it can discover/import markets, inspect positions, exit positions, modify persistent configuration, and perform live trades using additional credentials not fully declared in the top-level metadata. This mismatch can mislead users or orchestration systems into granting the skill more trust and access than intended, increasing the chance of unauthorized financial actions or inadvertent secret exposure.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
Not specified by scanner
Confidence
96% confidence
Finding

The manifest requests a SOLANA_PRIVATE_KEY even though the skill is described as a BTC/ETH correlation trader using Simmer/Kalshi-style markets, creating a strong mismatch between stated functionality and requested secrets. Unnecessary access to a blockchain private key materially increases the risk of secret theft or unauthorized signing if the skill code is compromised, misleading, or later updated.

Content

No source excerpt is available for this finding.

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
94% confidence
Finding

The skill declares access to sensitive environment variables such as SIMMER_API_KEY and SOLANA_PRIVATE_KEY but does not define any explicit tool scope or permission boundary. In an agent ecosystem, missing scope declarations reduce transparency and can allow broader-than-expected credential exposure or execution capability, which is especially risky for a trading skill handling live financial authority.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
98% confidence
Finding

The user-facing description says the skill requires only SIMMER_API_KEY and simmer-sdk, but the manifest also declares SOLANA_PRIVATE_KEY. This discrepancy can mislead users into granting a highly sensitive credential they did not expect to expose, undermining informed consent and making secret over-collection harder to detect.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

Declaring access to a private key without a clear warning or justification is dangerous because private keys enable irreversible financial actions and are far more sensitive than ordinary API tokens. In the context of a trading skill, undisclosed key access is especially risky because the skill could sign transactions or move assets without the user's informed understanding.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The file advertises SIMMER_API_KEY as the required credential, but the code supports a --live mode that executes real trades and the module docstring also notes a SOLANA_PRIVATE_KEY is needed for live trading. This mismatch can cause operators or automation to enable live trading without understanding that real-funds signing material is involved, increasing the chance of unsafe deployment and secret handling mistakes.

Content

No source excerpt is available for this finding.