The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructions direct the user to install the wribble package from the NPM registry. This is the official CLI tool required for the skill to function.
[COMMAND_EXECUTION]: The skill uses shell commands through the wribble binary to perform authentication and blog management. This is an expected use of a CLI-based tool wrapper.
[DATA_EXFILTRATION]: The skill is designed to upload local file contents and terminal command output to the Wribble blog service. While this involves sending data to an external server, it is the primary and documented purpose of the publishing skill.
[PROMPT_INJECTION]: The skill retrieves post content and channel data from an external API (Wribble). This data is ingested into the agent context without specific boundary markers or sanitization, creating a surface for indirect prompt injection where malicious instructions stored in a blog post could influence the agent.
Ingestion points: Data retrieved from wribble list, wribble get, and associated MCP tools.
Boundary markers: None identified; retrieved content is processed directly.
Capability inventory: Shell execution capabilities via the wribble tool.
Sanitization: None specified for external content processing.

wribble