The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill requires the installation of the @membranehq/cli package from npm, which is the official tool for the Membrane platform.
[COMMAND_EXECUTION]: The skill provides instructions for the agent to execute shell commands via the membrane CLI for logging in, connecting to services, and running data-retrieval actions.
[PROMPT_INJECTION]: The skill ingests untrusted data from external websites via the Ahrefs API, creating a surface for indirect prompt injection attacks.
Ingestion points: Data such as page titles, anchor text, and URL paths are retrieved through actions like get-backlinks, get-pages, and get-anchors (SKILL.md).
Boundary markers: There are no explicit delimiters or instructions provided to the agent to treat the retrieved SEO data as untrusted content.
Capability inventory: The agent can perform network requests and execute command-line operations using the membrane tool (SKILL.md).
Sanitization: No sanitization or content filtering is implemented for the external data before it is processed by the agent.

ahrefs