The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill connects to https://www.mixdao.world/api/latest to fetch data and https://api.minimaxi.com/anthropic to interact with an AI model. These connections are necessary for the skill's documented functionality.
[DATA_EXFILTRATION]: The skill requires the user to provide MIXDAO_API_KEY and ANTHROPIC_API_KEY via environment variables. It transmits the MIXDAO_API_KEY to www.mixdao.world and the ANTHROPIC_API_KEY to api.minimaxi.com. While this is standard for API-based tools, users should ensure they trust these endpoints before providing credentials.
[INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes untrusted data from an external API.
Ingestion points: Data is ingested from https://www.mixdao.world/api/latest in scripts/01-fetch.js.
Boundary markers: The LLM prompts in scripts/02-briefing.js (lines 197 and 248) do not include explicit delimiters or instructions to ignore potential commands embedded within the retrieved news items.
Capability inventory: The skill possesses write capabilities, using the patchMixdao function in scripts/02-briefing.js to submit data back to the mixdao.world API.
Sanitization: While the code uses JSON.stringify to structure data sent to the AI, it lacks semantic sanitization to prevent the LLM from obeying instructions hidden within the fetched content.

daily-briefing