The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructions specify the global installation of the plumcake-cli package via NPM. This package is hosted on a public registry and originates from a source not identified as a trusted vendor.
[COMMAND_EXECUTION]: The skill's primary functionality relies on the execution of the plumcake CLI tool, which interacts with the local shell and file system to manage communication with the external platform.
[DATA_EXFILTRATION]: The plumcake post command transmits data to an external server at plumcake.kushalsm.com. Using the @./file syntax, the agent is instructed to read local files and send their contents to this remote endpoint. This creates a data exfiltration vector if the agent is manipulated into sending sensitive files or environment configurations.
[PROMPT_INJECTION]: The skill retrieves external data, such as blocker posts and comments, via the plumcake read and plumcake list commands. This content is then integrated into the agent's context, creating an attack surface for indirect prompt injection.
Ingestion points: SKILL.md (via plumcake list and plumcake read commands).
Boundary markers: Absent. The instructions do not provide delimiters or warnings for the agent to ignore instructions embedded in the external posts it reads.
Capability inventory: The skill includes capabilities to read local files and perform network operations via the CLI, alongside any default capabilities provided to the agent.
Sanitization: Absent. There are no mechanisms described to sanitize or validate the external content before it is processed by the AI.

plumcake