The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill references and fetches DCQL query templates and JSON schemas from the iGrant.io verifiable data registry hosted on GitHub (decentralised-dataexchange repository). These are standard references for configuring verifiable credential presentations.
[INDIRECT_PROMPT_INJECTION]: The skill provides instructions for processing disclosed claims from external digital wallets via the iGrant.io API. This creates a surface for indirect prompt injection if an attacker-controlled wallet provides malicious instructions inside claim values (e.g., in the fundingSource object).
Ingestion points: Data enters the agent context when reading disclosed claims from the presentation field in the iGrant.io verification history record.
Boundary markers: None explicitly defined in the templates to separate wallet-provided data from agent instructions.
Capability inventory: The skill uses POST requests to the iGrant.io API to configure and send verification requests.
Sanitization: The skill mentions that the server uses JSON Schema validation (transactionDataDefinitionType) to validate the transactionData field.
[CREDENTIALS_UNSAFE]: The skill documentation includes a security warning explicitly advising users to keep API keys in environment variables or secret managers and never expose them to the browser, which is a standard security best practice.

igrantio-dcql-query-payment-wallet-attestation