The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The scripts/init.py script automatically executes a pip install command for the top-coder-ai-skills-debugger package. This dependency is not from a trusted vendor list and is installed without version pinning, posing a risk of supply chain attack if the package is compromised.
[COMMAND_EXECUTION]: The scripts/debug.py script enables the execution of Node.js programs and supports an evaluate action. This allows the agent to run arbitrary JavaScript code within the debugged environment, which could be exploited to perform unauthorized operations if the agent is manipulated.
[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by ingesting and processing runtime data that could be controlled by an attacker.
Ingestion points: The skill reads and returns stack traces, local variables, and evaluation results to the agent via scripts/debug.py.
Boundary markers: No explicit delimiters or safety instructions are used to separate debugger output from system instructions.
Capability inventory: The agent context associated with this skill likely possesses file system and network access, which could be targeted via instructions embedded in debugged data.
Sanitization: There is no evidence of sanitization or filtering of the content retrieved from the Node.js runtime environment.

debugger-nodejs