The Agent Skills Directory

[COMMAND_EXECUTION] (HIGH): The skill instructs the agent to execute shell commands using npx -y bun to run local TypeScript scripts. This allows the agent to spawn subprocesses and interact with the system shell directly to perform automation tasks.
[CREDENTIALS_UNSAFE] (HIGH): The skill manages and persists X (Twitter) browser sessions. It explicitly uses Chrome profiles to bypass anti-automation, meaning the agent has access to sensitive session cookies and authentication tokens stored in the browser profile directories (--profile <dir>).
[PROMPT_INJECTION] (HIGH): Category 8 (Indirect Prompt Injection). The skill is designed to process untrusted external content (text, Markdown articles) and post it to a public platform.
Ingestion points: Positional arguments for post text/comments and input file paths for Markdown articles (article.md).
Boundary markers: None identified. There are no delimiters or instructions to ignore embedded commands in the input data.
Capability inventory: Browser automation via CDP, shell command execution, and file system read access.
Sanitization: None identified. The skill lacks evidence of escaping or validating external content before processing.
[EXTERNAL_DOWNLOADS] (MEDIUM): Use of npx -y indicates that packages may be downloaded and executed at runtime. While targeting local scripts, the underlying dependencies of those scripts are not pinned or verified in the provided metadata.

baoyu-post-to-x