The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill provides instructions for downloading the Pensar Apex CLI from the vendor's domain (pensarai.com), as well as via Homebrew and a global NPM package (@pensar/apex).\n- [REMOTE_CODE_EXECUTION]: The skill provides setup commands that download and execute scripts from the internet, such as curl -fsSL https://pensarai.com/install.sh | bash for Linux/macOS and irm https://www.pensarai.com/apex.ps1 | iex for Windows. These are presented as official installation methods for the tool.\n- [COMMAND_EXECUTION]: The skill facilitates the execution of various pensar CLI commands for autonomous penetration testing, threat modeling, and managing workspace attack surface data.\n- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from target URLs and local source code during analysis, which represents a surface for indirect prompt injection.\n
Ingestion points: Target URLs provided via --target, local source code directories via --cwd, and user-provided objectives via --objective.\n
Boundary markers: No delimiters or instructions are specified to separate external content from the agent's instructions.\n
Capability inventory: The agent can execute CLI commands, perform network scanning, and read/write files in the ~/.pensar/ directory.\n
Sanitization: No specific sanitization or filtering of target-derived data is defined in the skill instructions.

pensar-security