The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill instructions include a command to download and execute a script from a remote URL by piping it directly to the bash shell.
Evidence: curl -fsSL https://tempo.xyz/install | bash found in the 'One-time setup' section of SKILL.md.
[EXTERNAL_DOWNLOADS]: The skill relies on a third-party service and CLI tool (tempo) hosted at tempo.xyz, which is outside of verified primary infrastructure providers.
Evidence: The skill requires tempo as an external binary and directs users to https://docs.tempo.xyz for installation and documentation.
[COMMAND_EXECUTION]: The instructions guide the agent or user to execute commands that manage sensitive operations, including wallet authentication and funding.
Evidence: Commands such as tempo wallet login and tempo wallet fund are presented as necessary setup steps.
[INDIRECT_PROMPT_INJECTION]: The skill creates a vulnerability surface by ingesting and potentially acting upon untrusted data from external API responses.
Ingestion points: Data returned from tempo request calls to api.nansen.ai (e.g., netflow data, holder lists).
Boundary markers: The instructions lack explicit delimiters or warnings to ignore instructions embedded in the API responses.
Capability inventory: The skill has the capability to execute shell commands and perform network requests.
Sanitization: No sanitization or validation of the API response content is described before the agent processes the data.

nansen-mpp-payment