The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes shell commands using the Agentspace CLI (ascli) and package managers like npm or pnpm to perform folder synchronization and sharing operations.
[EXTERNAL_DOWNLOADS]: The skill facilitates the installation and execution of the @agentspace-so/ascli package from the npm registry, which is the official tool provided by the vendor.
[DATA_EXFILTRATION]: The core functionality involves uploading local data to agentspace.so. This is the intended purpose of the skill, and the destination matches the vendor's domain. The skill instructions mandate that the agent must not assume the target path and must confirm the file or folder location with the user before proceeding, which serves as a safeguard against accidental data exposure.
[PROMPT_INJECTION]: The skill has an indirect prompt injection surface because it processes local files that could contain malicious instructions. Evidence:
Ingestion points: Any local file or directory the user instructs the agent to share (SKILL.md).
Boundary markers: Not present for the content of the files being uploaded.
Capability inventory: Shell command execution via ascli, npm, and pnpm (SKILL.md, references/commands.md).
Sanitization: The skill requires explicit path confirmation to prevent accidental sharing but does not analyze or sanitize the content of the files themselves.

agentspace