The Agent Skills Directory

[SAFE]: The skill implements strong security boundaries by requiring backend verification for all transactions and explicitly forbidding the granting of entitlements based on unverified client-side assertions or pending states. It also provides guidance on sensitive data handling, advising against the logging or exposure of purchase tokens and receipts. Instructions include robust logic for handling account transitions and session lifecycle to prevent entitlement leakage between users.- [INDIRECT_PROMPT_INJECTION]: The skill processes external transaction data from app stores and backend verification results. Ingestion points include store purchase listeners (referenced in references/purchase-lifecycle-and-entitlements.md) and configuration files like pubspec.yaml (referenced in SKILL.md). The skill mitigates risks by requiring normalization, rejecting malformed or unverified data, and enforcing a trusted backend verification boundary before granting entitlements. Capabilities are strictly limited to billing state reconciliation and entitlement management based on authoritative store sources.

flutter-in-app-purchases