The Agent Skills Directory

[DATA_EXFILTRATION]: The skill manages sensitive information by reading and writing a Hiro API key to a local configuration file at ~/.aibtc/config.json.
Evidence: Subcommands set-hiro-api-key, get-hiro-api-key, and delete-hiro-api-key in settings.ts perform file-based storage operations for these credentials.
Context: This is the primary function of the skill and follows standard practices for CLI configuration management.
[EXTERNAL_DOWNLOADS]: The skill connects to external endpoints to perform health diagnostics and version checks.
Evidence: settings.ts fetches data from the npm registry (registry.npmjs.org) and the Hiro API (api.hiro.so), which are well-known services.
Evidence: settings.ts also connects to the vendor-owned relay health endpoint at sponsor.aibtc.dev.
[PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection by ingesting JSON data from external web services and outputting it to the agent's context.
Ingestion points: settings.ts fetches content from external APIs including npm, Hiro, and a user-definable relay URL.
Boundary markers: The skill does not use specific delimiters or instructions to ignore potential commands within the fetched JSON data.
Capability inventory: The skill has the ability to read and write to the local configuration file ~/.aibtc/config.json.
Sanitization: Data from external fetches is parsed as JSON but not otherwise sanitized before being presented to the agent.

settings