The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes a suite of local Python and Bash scripts to manage document catalogues and client overrides. These scripts perform standard operations using system tools like curl, openssl, and awk to interact with the Sumsub API.
[EXTERNAL_DOWNLOADS]: Communicates with official Sumsub API endpoints (e.g., api.sumsub.com) to retrieve document configuration, client settings, and entitlements. These operations are essential for the skill's stated purpose of managing identity document settings.
[SAFE]: The skill implements significant safety guardrails, including a sandbox-only execution mode for API tokens (unless explicitly overridden) and hardcoded restrictions in build_country_mappings.py that prevent the agent from modifying security-sensitive fields like shouldBeSigned and shouldBeMasked.
[SAFE]: Data processing is performed using safe parsing methods, such as yaml.safe_load() and standard json.load(). The skill also provides explicit warnings to the user when modifying critical fraud-prevention settings like acceptScreenshots or acceptDigitalDoc.
[INDIRECT_PROMPT_INJECTION]: The skill ingests data from the Sumsub API (global settings and catalogues) to populate its context. It mitigates injection risks by using structured JSON processing and local script-based filtering rather than directly interpolating large blocks of external text into the prompt.

sumsub-supported-id-documents