The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill installs the @aisuite/chub package from the NPM registry to provide the necessary command-line functionality.
[PRIVILEGE_ESCALATION]: The skill uses npm install -g, which is a privileged operation that typically requires administrative or root access on many operating systems.
[INDIRECT_PROMPT_INJECTION]: The skill is instructed to treat content fetched via the chub utility as the primary "authority" and "source of truth" for workflows, model names, and configurations. This creates a risk where malicious or compromised content in the external repository could misdirect the agent.
Ingestion points: Content is ingested into the agent context via the output of chub search and chub get commands.
Boundary markers: Absent; there are no instructions to the agent to isolate or treat the fetched content as untrusted data.
Capability inventory: The skill allows for shell command execution, global package installation, and network-based feedback transmission.
Sanitization: Absent; the skill does not define any validation or filtering for the data retrieved from the remote service.
[DATA_EXFILTRATION]: The skill includes a chub feedback command designed to send descriptive text and labels back to a remote server. While the documentation provides warnings against including sensitive information, the functionality serves as an outbound data channel.
[COMMAND_EXECUTION]: The skill executes multiple subcommands through the system shell, including command -v for environment checks and several chub CLI operations for data retrieval and feedback.

context-hub