Back to skill

Security audit

Indicator Lab

Security checks for vulnerabilities and agentic risk

Overview

This skill appears to be a simple finance backtesting guide with expected network/file permissions, but its referenced script is missing and its dependency install is not pinned.

Install only in a dedicated virtual environment, avoid elevated privileges, and consider pinning or reviewing dependency versions before use. Also verify the missing backtest script is supplied elsewhere before expecting the skill to run.

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
SKILL.md:25
Finding

Unpinned Third-Party Dependencies Enable Supply-Chain Compromise

Content
View full analysis

Vulnerability Details

File Location: SKILL.md:25-27 and instructions.md:25-27
Vulnerability Type: Unpinned and unverifiable third-party dependencies
Risk Level: Medium

Complete vulnerable code snippet:

text
## Requirements
bash
pip install pandas pandas-ta yfinance plotly

The same installation instructions appear in both affected files.

Technical Analysis

The documented installation command retrieves four packages without pinned versions, cryptographic hashes, a lockfile, or an explicitly trusted package index. Package resolution therefore depends on mutable package-index state and the user's local pip configuration.

Python package installation can execute package-controlled build code. If a package release, transitive dependency, configured package index, or similarly named package is compromised, malicious code could execute during installation. Because no hashes are specified, pip cannot verify that users receive artifacts reviewed by the project publisher.

This also prevents reproducible installation: two users running the same command at different times may receive materially different dependency versions and code.

Attack Path

  1. An attacker compromises a listed package, one of its transitive dependencies, or a package source configured in the victim's pip environment.
  2. The attacker publishes a malicious distribution or modifies a package artifact that remains compatible with the unconstrained dependency request.
  3. A user follows the documented command:
    bash
    pip install pandas pandas-ta yfinance plotly
    
  4. pip resolves and downloads the attacker-controlled package because no approved versions or hashes constrain resolution.
  5. Malicious build or installation logic executes with the privileges of the user running pip.
  6. The installed package can continue executing when imported by downstream Python code.

Impact Assessmen

...[truncated 714 chars]

Remediation
View remediation

Remediation Suggestions

  1. Replace the free-form installation command with a reviewed requirements or lock file containing exact versions.
  2. Generate and record SHA-256 hashes for every direct and transitive dependency, then install with hash verification:
    bash
    python -m pip install --require-hashes -r requirements.txt
    
  3. Use a dependency-management process that resolves and locks transitive dependencies, such as pip-tools, Poetry, or an equivalent reproducible workflow.
  4. Explicitly document the trusted package index and prevent unintended fallback to untrusted indexes.
  5. Install dependencies in a dedicated, non-privileged virtual environment rather than globally or as an administrative user.
  6. Review locked dependency artifacts and use automated vulnerability and provenance scanning before publishing updates.
  7. Apply the corrected dependency instructions consistently to both SKILL.md and instructions.md.
Vulnerability Patterns
  • 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
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep