The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses subprocess.run in a Python script and npx commands in the instructions to execute various external security tools like nmap, subfinder, and nuclei. This behavior is essential to the skill's purpose of automating security assessments.
[EXTERNAL_DOWNLOADS]: The instructions include running npx @keygraph/shannon@latest, which fetches and executes code from the NPM registry. This is a standard mechanism for utilizing developer and security CLI tools.
[INDIRECT_PROMPT_INJECTION]: The provided Python script ai_pentester.py ingests data from reconnaissance tools directly into LLM prompts to guide subsequent exploitation phases. While this presents an attack surface where malicious tool output could influence the agent's actions, the script mitigates this by using internal mapping logic to translate LLM decisions into specific tool handlers.

ai-pentesting