The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill provides instructions to download and execute scripts directly from an external domain using shell pipes, which is a high-risk pattern for executing unverified code.
Evidence: curl -fsSL https://d.officecli.ai/install.sh | bash in SKILL.md.
Evidence: irm https://d.officecli.ai/install.ps1 | iex in SKILL.md.
[EXTERNAL_DOWNLOADS]: The skill attempts to download binaries and scripts from officecli.ai, a domain not recognized as a trusted organization or well-known service.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data (Office documents) and has broad execution capabilities, creating a significant attack surface.
Ingestion points: The agent reads content from .docx, .xlsx, and .pptx files using commands like view, get, query, and raw (all in SKILL.md).
Boundary markers: The instructions lack delimiters or warnings to ignore embedded instructions within the processed documents.
Capability inventory: The skill uses a CLI tool to modify the filesystem, execute shell commands, and potentially interact with the network during installation.
Sanitization: There is no mention of sanitizing or validating the contents of the Office documents before processing them.
[COMMAND_EXECUTION]: The core functionality of the skill involves executing the officecli binary and other shell commands to manipulate files and document structures.

officecli