The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill requires the user to install the xactions package from the npm registry using the command npm install -g xactions.
[COMMAND_EXECUTION]: The skill documentation lists various CLI commands that execute locally to scrape data and manage configurations, including the mcp-config --write command which modifies local IDE configuration files.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external web sources (X/Twitter) that is then incorporated into the agent's context.
Ingestion points: Data is retrieved from the web via commands such as xactions search, xactions tweets, xactions followers, and xactions profile.
Boundary markers: The skill does not define specific delimiters or instructions for the agent to distinguish between its own logic and the content of the scraped tweets or profiles.
Capability inventory: The skill has the ability to write scraped data to local files in JSON or CSV format and can write configuration data directly to system paths for various clients (Claude Desktop, VS Code, etc.).
Sanitization: No mention of sanitization or filtering of the scraped content is provided to prevent the agent from interpreting embedded instructions within the data.
[METADATA_POISONING]: The author handle specified in the YAML metadata ('nichxbt') does not match the attributed author name ('nirholas').

xactions-cli