T08 · Insecure Dependencies
- Location
SKILL.md:22- Finding
Unpinned Third-Party npm Dependency
- Content
View full analysis
- Remediation
View remediation
Security audit
Security checks for vulnerabilities and agentic risk
This Yahoo Finance skill is mostly purpose-aligned, but its install instructions can persistently modify a system-wide command path and use an unpinned npm package.
Review the install steps before using this skill. Prefer a user-local or project-local install, pin and audit yahoo-finance2, avoid sudo and forced /usr/local/bin symlinks unless you have verified the target binary and existing yf command, and understand that yahoo-finance2 may store cookies in ~/.yf2-cookies.json.
SKILL.md:22Unpinned Third-Party npm Dependency
SKILL.md:4Forced Replacement of a System-Wide Tool Link
Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).
rm -f ~/.yf2-cookies.json
yf quote AAPL
The activation description says the skill should be used when the user asks to "get financial data" or "needs stock market data," which are broad natural-language triggers without clear boundaries or exclusion conditions. This increases the chance of unintended invocation for generic finance questions that may not specifically call for this CLI-based Yahoo Finance skill.
The documented use of sudo to create a symlink in /usr/local/bin requires elevated privileges and modifies a globally trusted executable path. Even though the command appears intended for setup, encouraging root-level execution in a skill increases the risk of privilege misuse, path hijacking, or unintended system modification if the binary path is incorrect or attacker-controlled.
brew install jq
npm install yahoo-finance2
sudo ln -s /opt/homebrew/bin/yahoo-finance /usr/local/bin/yf
The skill includes installation and troubleshooting commands that make system changes, including privileged symlink creation and deletion of a local file, without explicit warnings or safer alternatives. In an agent context, normalizing such commands can lead users to perform risky actions they do not fully understand, potentially altering trusted paths or deleting the wrong file if adapted carelessly.
No suspicious patterns detected.