The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill requires the installation of third-party Python packages twikit and browser_cookie3 from the Python Package Index (PyPI) to facilitate X GraphQL interaction and browser cookie retrieval.
[COMMAND_EXECUTION]: The skill executes shell commands to install dependencies via pip and runs a local Python script scripts/xfetch.py to perform data extraction tasks.
[DATA_EXFILTRATION]: The skill contains functionality (auth extract command) to access local browser profile directories to retrieve session cookies for x.com. It stores these credentials in a local configuration file at ~/.config/xfetch/cookies.json using restricted file permissions (0o600), which is a standard and documented feature of the tool's primary purpose.
[INDIRECT_PROMPT_INJECTION]: The skill presents a surface for indirect prompt injection by ingesting unmoderated user content from X.com.
Ingestion points: Untrusted data enters the agent context via GraphQL API responses processed in scripts/xfetch.py (specifically in the extract_timeline and parse_tweet functions).
Boundary markers: The skill does not employ specific delimiters or instructions to the agent to ignore potentially malicious content embedded within the fetched tweets.
Capability inventory: The skill has the capability to perform network requests to X.com, write to the local filesystem (JSON, CSV, SQLite), and read local browser data.
Sanitization: The script does not perform sanitization of the retrieved tweet text before outputting it to the agent.

xfetch