The Agent Skills Directory

[SAFE]: Analysis of the skill instructions, metadata, and provided Motoko code templates reveals no evidence of malicious patterns such as prompt injection, obfuscation, or unauthorized persistence mechanisms.- [EXTERNAL_DOWNLOADS]: The skill utilizes the x-client package (version ~0.3.0) via the Motoko Package Manager (mops). This is a standard and expected dependency for the Internet Computer platform used for its documented purpose of interfacing with the X API v2.- [INDIRECT_PROMPT_INJECTION]: The skill establishes a data ingestion surface by allowing user-provided content to be posted to an external social media platform. While inherent to the skill's purpose, the following factors are noted:
Ingestion points: The accessToken and body parameters in the postTweet function within src/backend/main.mo.
Boundary markers: None present in the provided code snippet.
Capability inventory: Performs network operations via TweetsApi.createPosts to the X API.
Sanitization: No explicit sanitization or filtering of the tweet body is performed in the template.

connector-x