The Agent Skills Directory

[EXTERNAL_DOWNLOADS] (SAFE): The skill references standard NPM packages bakong-khqr and ts-khqr. These are relevant to the primary purpose of KHQR integration and do not use suspicious installation methods like piped shell scripts.
[DATA_EXFILTRATION] (SAFE): Communication is directed to official Bakong infrastructure (api-bakong.nbc.org.kh). There are no patterns suggesting exfiltration of sensitive files or credentials to unknown domains.
[PROMPT_INJECTION] (SAFE): No instructions were found that attempt to subvert the agent's behavior, extract system prompts, or bypass safety protocols.
[INDIRECT_PROMPT_INJECTION] (LOW): The skill defines a surface for processing untrusted data (KHQR strings and API responses). Although this presents a theoretical attack surface, it is consistent with the skill's utility and the risk is mitigated by recommended validation steps.
Ingestion points: references/khqr-sdk.md (QR decoding) and references/open-api.md (API response processing).
Boundary markers: Not explicitly defined in the reference snippets, though the data is handled as structured JSON.
Capability inventory: Includes network operations for transaction status checks.
Sanitization: The documentation explicitly recommends using BakongKHQR.verify to validate inputs before processing.

bakong-khqr