The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill uses the npm install command to fetch the @membranehq/cli package. This is the official tool for the Membrane platform, facilitating secure communication with external services.
[COMMAND_EXECUTION]: The integration relies on executing shell commands via the membrane CLI to authenticate users and perform accounting tasks like listing invoices or creating contacts. This is the primary mechanism for the skill's functionality.
[INDIRECT_PROMPT_INJECTION]: The skill processes external data from Xero, which presents an attack surface where malicious instructions could be embedded in accounting records.
Ingestion points: Data from Xero is retrieved via membrane action run and proxy membrane request commands (found in SKILL.md).
Boundary markers: There are no specific instructions for the agent to use delimiters or ignore embedded instructions within the Xero data.
Capability inventory: The skill uses the membrane CLI (SKILL.md) which has the ability to read and write sensitive accounting data.
Sanitization: No explicit validation or sanitization steps are defined for the data returned from Xero APIs before it is processed by the agent.

xero