The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the agent to install the x-twitter-pp-cli and x-twitter-pp-mcp tools from the author's official GitHub repository (github.com/mvanhorn/printing-press-library) and npm organization (@mvanhorn/printing-press-library). These are documented as the primary tools for the skill's operation.
[COMMAND_EXECUTION]: The skill requires the execution of shell commands for dependency installation (npx, go install) and operational tasks via the custom CLI binary. These commands are necessary for the skill's stated purpose of managing X data.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted content from the X (Twitter) API, including posts, threads, and bookmarks, which may contain adversarial instructions.
Ingestion points: External data enters the agent context through search, thread reconstruction, and bookmark retrieval commands in SKILL.md.
Boundary markers: The CLI provides a --agent flag that enforces structured JSON output, helping to separate data from instructions.
Capability inventory: The CLI is capable of writing to a local SQLite database, exporting files in various formats, and delivering output to external webhooks.
Sanitization: The skill supports the --select flag to limit the data fields returned to the agent, reducing the overall surface for injection attacks.

pp-x-twitter