The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill describes interactions with external API endpoints, which could allow maliciously crafted data from those endpoints to influence agent behavior.
Ingestion points: The agent is instructed to fetch and process data from endpoints like /v2/config/digital-wallet/openid/key-management/keys in SKILL.md.
Boundary markers: No specific delimiters or boundary instructions are provided to separate API data from agent instructions.
Capability inventory: The skill involves sensitive operations such as key creation, vault configuration, and CSR generation.
Sanitization: No instructions for sanitizing or validating the contents of the API responses are included.
[CREDENTIALS_UNSAFE]: The documentation notes that certain API responses from the iGrant.io service include sensitive credentials (e.g., vaultPassword and clientSecret) in cleartext. The skill provides an explicit warning to protect this data, consistent with the vendor's platform documentation.

igrantio-api-key-management