The Agent Skills Directory

[EXTERNAL_DOWNLOADS] (MEDIUM): The skill prompts the installation of @miketromba/polar-cli via npm. Neither the package nor the author is within the trusted scope, posing a risk of supply chain attack or malicious code execution at runtime.
[COMMAND_EXECUTION] (MEDIUM): The skill is designed to execute shell commands that interact with SaaS billing and resource management. This includes destructive operations (e.g., polar customers delete --yes) which, while functional, increase the impact of any successful injection attack.
[CREDENTIALS_UNSAFE] (LOW): The documentation guides the user to provide a POLAR_ACCESS_TOKEN. While it correctly recommends environment variables, the agent's involvement in handling these high-privilege credentials requires careful monitoring.
[PROMPT_INJECTION] (HIGH): The skill is highly vulnerable to Indirect Prompt Injection (Category 8). It instructs the agent to fetch documentation from https://polar.sh/docs/llms.txt and use it to "understand how a feature works" or "verify API behavior." This remote content is untrusted data that enters the agent's context.
Ingestion points: Remote text index and linked documentation pages at polar.sh.
Boundary markers: Absent. The agent is not instructed to disregard embedded commands or instructions within the documentation.
Capability inventory: The agent can create/delete customers, manage billing, and issue license keys via the polar CLI.
Sanitization: Absent. The skill lacks any validation or filtering of the fetched documentation before the agent interprets it for decision-making.

polar