The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill fetches an OpenAPI schema configuration from a remote GitHub repository hosted at https://raw.githubusercontent.com/holon-run/uxc/main/skills/blockscout-openapi-skill/references/blockscout-v2.openapi.json. Additionally, the skill interacts with external Blockscout API endpoints, one of which (https://optimism.blockscout.com/api/v2) has been flagged by automated scanners.
[COMMAND_EXECUTION]: The skill instructs the agent to use the uxc CLI tool to link and execute commands, such as uxc link blockscout-openapi-cli and blockscout-openapi-cli get:.... This involves shell execution for network-bound operations.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external blockchain explorer APIs, creating a potential surface for indirect prompt injection if the data contains malicious instructions targeted at the agent.
Ingestion points: The skill reads transaction history, token metadata, and address summaries via the Blockscout API as defined in references/blockscout-v2.openapi.json.
Boundary markers: Instructions suggest parsing stable JSON fields (ok, kind, data), but there are no explicit warnings to ignore instructions embedded within the retrieved data.
Capability inventory: The skill utilizes uxc for network requests and command execution.
Sanitization: No evidence of data sanitization or escaping of the external API responses before interpolation into the agent's context is provided.

blockscout-openapi-skill