The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill's installation instructions include downloading scripts from a remote domain and piping them directly to shell interpreters (e.g., curl -fsSL https://d.officecli.ai/install.sh | bash and irm https://d.officecli.ai/install.ps1 | iex). This practice bypasses inspection of the code before execution and is a high-risk method for distributing software.
[COMMAND_EXECUTION]: The skill relies on executing a local binary (officecli) to perform complex operations on the file system and Office documents.
[INDIRECT_PROMPT_INJECTION]: The skill enables the agent to parse external document content, making it vulnerable to malicious instructions embedded in those files. 1. Ingestion points: Document content is read into the agent's context through commands like officecli get, officecli view text, and officecli query (SKILL.md). 2. Boundary markers: The skill does not provide any delimiters or instructions for the agent to ignore or isolate instructions found within the document data. 3. Capability inventory: The agent has the power to create, modify, and delete elements in files using set, add, remove, batch, and raw-set (SKILL.md). 4. Sanitization: There is no evidence of sanitization or filtering of the content extracted from documents before the agent processes it.

officecli