The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill has a significant attack surface for indirect prompt injection.
Ingestion points: User-provided parameters such as sandbox IDs, merchant numbers, and transaction details enter the context via scripts/render_server_example.sh.
Boundary markers: The skill lacks explicit delimiters or instructions to ignore embedded commands within the user-supplied data used for code generation.
Capability inventory: The skill can write files and execute local scripts to render project structures and process certificates.
Sanitization: There is no visible sanitization of user inputs before they are interpolated into generated code and configuration templates.
[EXTERNAL_DOWNLOADS]: The skill includes a utility script (scripts/client-sdk-scripts/download_sdk.sh) that fetches ZIP files from arbitrary, user-provided URLs using curl. While intended for downloading legitimate SDKs, this mechanism relies on the integrity of the user-supplied source.
[COMMAND_EXECUTION]: Several automation scripts (scripts/cert_to_base64.sh, scripts/render_server_example.sh) execute shell commands and system utilities like openssl. These scripts process user-provided files (certificates, configuration data) locally.
[OBFUSCATION]: Static analysis identified large hex blobs in assets/server-examples/python-quickstart/src/aipay_demo/utils/pfx.py and assets/server-examples/nodejs-quickstart/src/utils/pfx.js. Upon review, these are verified as legitimate lookup tables (S-boxes) for the RC2 cryptographic algorithm implementation.

jd-aipay-onboarding