Back to skill

Security audit

Polymarket Clob Microstructure

Security checks for vulnerabilities and agentic risk

Overview

This is a disclosed trading scanner that defaults to dry-run, with live trading gated by an explicit flag and API key.

Install only if you intend to let this skill analyze prediction markets and potentially trade when run with --live. Use a dedicated Simmer API key with limited permissions and spending limits, keep dry-run mode until you trust the strategy, and prefer pinned dependency versions before unattended live use.

Vulnerability Patterns
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • 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 (1)

T08 · Insecure Dependencies

Warning
Location
clawhub.json:4
Finding

Unpinned Executable Third-Party Dependencies

Content
View full analysis

Vulnerability Details

File Location: clawhub.json, lines 4–8
Vulnerability Type: Unpinned third-party dependencies
Risk Level: Medium

json
"requires": {
  "pip": [
    "simmer-sdk",
    "requests"
  ],

Technical Analysis

The project installs simmer-sdk and requests without exact version constraints or integrity hashes. Consequently, future installations may retrieve different code than the versions originally reviewed.

This is particularly sensitive because simmer-sdk is imported and executed by the Skill, receives the SIMMER_API_KEY, communicates with the trading service, and can submit financial trades when live mode is enabled. A compromised package release, compromised maintainer account, malicious dependency update, or altered transitive dependency could therefore execute code with the same process privileges as the Skill.

The audit found no evidence that either currently named package is malicious. The vulnerability is the absence of dependency version and integrity controls.

Attack Path

  1. An attacker compromises an upstream package, its publishing account, or one of its transitive dependencies.
  2. The attacker publishes a malicious version under the legitimate package name.
  3. A managed installation or update resolves the unpinned dependency to that malicious version.
  4. Python imports and executes the dependency when clob_microstructure.py starts.
  5. The malicious package reads process-accessible data such as SIMMER_API_KEY, intercepts SDK requests, modifies returned market information, or invokes trading operations.
  6. The attacker exfiltrates the credential or causes unauthorized trades within the API key's permissions.

Impact Assessment

Successful exploitation would grant the malicious dependency the privileges of the Skill process. This includes access to environment variables, including SIMMER_API_KEY; local files available to the process; netwo ...[truncated 303 chars]

Remediation
View remediation

Remediation Suggestions

  1. Pin each direct dependency to an exact, reviewed version, for example package==x.y.z.
  2. Generate a lock file that also fixes all transitive dependency versions.
  3. Require cryptographic hashes during installation, such as a hash-locked requirements file used with pip --require-hashes.
  4. Review package provenance, release history, maintainers, and artifacts before approving upgrades.
  5. Perform dependency vulnerability and malware scanning in CI and fail builds on unreviewed lock-file changes.
  6. Install from a controlled package index or approved artifact repository where feasible.
  7. Run the Skill under a dedicated, minimally privileged account with restricted filesystem and network access.
  8. Use a dedicated Simmer API key with minimum trading permissions, explicit spending limits, monitoring, and rapid revocation capability.
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (1)

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

This skill can place real trades whenever it is invoked with the --live flag, but it does not require a second explicit confirmation from the user immediately before order submission. In an automated-agent context, accidental invocation, prompt-induced tool use, or misconfiguration could trigger irreversible financial actions, making the absence of a confirmation barrier materially risky.

Content

No source excerpt is available for this finding.