The Agent Skills Directory

[SAFE]: The skill provides detailed and secure integration guidance for Midtrans payment products (Snap, Core API, BI-SNAP) following industry standards for payment processing.
[SAFE]: Utility scripts in the scripts/ directory are provided for deterministic operations like signature generation, verification, and webhook replaying to local handlers, facilitating safe development workflows.
[SAFE]: The scripts implement robust safety features, such as the replay_snap_webhook.sh script which refuses to operate with production-prefixed keys unless an explicit --allow-production flag is passed, and restricts network targets to local environments by default.
[SAFE]: The skill emphasizes the critical importance of redacting secrets, authentication tokens, and customer PII in logs and debugging output across all documentation and scripts.
[SAFE]: Use of well-known external dependencies. The skill utilizes the standard cryptography Python package for signature operations and openssl for cryptographic hashes.
[SAFE]: Indirect Prompt Injection Surface. The skill processes transaction data and webhook payloads (as demonstrated in assets/fixtures/), but mitigates risks through strict signature verification requirements and clear sanitization guidance.
Ingestion points: JSON fixtures in assets/fixtures/ and payloads passed via CLI flags.
Boundary markers: Emphasizes mandatory signature verification before any order mutation.
Capability inventory: Scripts for signature verification and webhook replaying.
Sanitization: Comprehensive redaction rules for logging business identifiers only.

integrate-midtrans-payments