The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill uses twitter-cli to automatically extract session cookies from browser profiles (Chrome, Arc, Edge, Firefox, Brave) to authenticate with Twitter/X without API keys. It also encourages setting environment variables like TWITTER_AUTH_TOKEN and TWITTER_CT0 which contain sensitive session data.\n- [COMMAND_EXECUTION]: The documentation suggests running security unlock-keychain ~/Library/Keychains/login.keychain-db on macOS. This command allows the CLI tool to decrypt cookies but also lowers the security posture of the system keychain.\n- [EXTERNAL_DOWNLOADS]: The skill's setup process involves installing twitter-cli from an external source via uv tool install twitter-cli. This introduces a dependency on a third-party tool (jackwener/twitter-cli) that is not part of the trusted vendor list.\n- [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection as it processes untrusted data from Twitter (tweets, search results, and profiles).\n
Ingestion points: Data enters through twitter feed, twitter search, and twitter user-posts commands in SKILL.md.\n
Boundary markers: No explicit delimiters or instructions to ignore embedded commands in the fetched data are present.\n
Capability inventory: The agent can execute system commands via twitter-cli, write output to files using -o, and perform network operations to x.com.\n
Sanitization: There is no evidence of sanitization or filtering of the fetched tweet content before it is presented to the agent for analysis.

twitter