feishu-openapi-skill
Pass
Audited by Gen Agent Trust Hub on Oct 16, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill fetches its curated OpenAPI schema from the author's official GitHub repository (holon-run/uxc). This is a standard vendor resource used to define the API interface for the uxc tool.
- [COMMAND_EXECUTION]: The skill utilizes several local command-line tools including
uxc,curl,jq, andrg. These are used appropriately for executing API requests, performing authentication bootstrapping, and running validation checks on the skill files. - [INDIRECT_PROMPT_INJECTION]: The skill processes external data such as chat history and incoming messages. This presents a potential surface for indirect prompt injection. The skill includes guardrails recommending the use of JSON output envelopes instead of raw text to prevent the agent from misinterpreting external content as instructions.
- Ingestion points: Chat history via
get:/im/v1/messagesand chat details viaget:/im/v1/chats. - Boundary markers: Documentation advises keeping automation on the JSON output envelope to ensure the agent parses structured fields rather than raw text.
- Capability inventory: Subprocess calls via
uxc, file-writes via the subscription sink, and network operations to Feishu/Lark endpoints. - Sanitization: Relies on structured OpenAPI schema validation and the
uxctool's handling of API responses.
Audit Metadata