The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted content from X (Twitter) through search results, timelines, and user profiles. This represents a surface for indirect prompt injection where malicious tweets could attempt to influence agent behavior.
Ingestion points: The script x.py retrieves external data in cmd_search, cmd_timeline, cmd_user_tweets, and cmd_tweet.
Boundary markers: No specific delimiters are used in the prompt instructions to isolate external content, though the script returns structured JSON.
Capability inventory: The skill possesses write capabilities including posting, deleting, and following, which are correctly gated by a requirement for explicit user confirmation.
Sanitization: The script formats data into JSON and truncates tweet text, reducing the risk of large-scale injection payloads.
[COMMAND_EXECUTION]: The skill uses the Bash tool to execute a local Python script x.py. This is the intended mechanism for the skill's functionality and is used securely with fixed script paths.
[EXTERNAL_DOWNLOADS]: The skill communicates with X (Twitter) API endpoints and official asset domains (abs.twimg.com). These network operations are necessary for the skill's core purpose of interacting with the service.
[CREDENTIALS_SAFE]: The skill handles sensitive account cookies (X_COOKIES). It follows security best practices by explicitly instructing the agent never to print or echo these secrets in its output.

x