The Agent Skills Directory

[SAFE]: The skill's functionality aligns with its stated purpose of providing access to the Felo SuperAgent API. All identified external resources and domains belong to the official vendor, felo-inc.
[EXTERNAL_DOWNLOADS]: The skill's documentation describes an installation process using npx to fetch content from the vendor's official GitHub organization (Felo-Inc). This is a standard deployment method for this platform.
[COMMAND_EXECUTION]: The skill executes a bundled JavaScript file (run_superagent.mjs) via Node.js to manage the streaming connection. This is the intended execution model for the skill's features.
[DATA_EXFILTRATION]: Network communication is directed to the official API endpoint https://openapi.felo.ai. This is required for the skill to function and is documented for the user.
[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface as it processes user input and external AI responses.
Ingestion points: The USER_QUERY_HERE variable in SKILL.md receives untrusted data from the user session.
Boundary markers: User input is passed as a command-line argument within quotes, though no explicit delimiters are used to separate user instructions from system context in the final script execution.
Capability inventory: The skill uses a bash tool to run local scripts and performs network requests to the vendor's backend.
Sanitization: No explicit sanitization or filtering logic for external content is provided in the analyzed markdown files.

felo-superAgent