The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes data from Twitter/X, which is an untrusted external source, creating a potential vector for indirect prompt injection.
Ingestion points: Tools such as TWITTER_RECENT_SEARCH, TWITTER_POST_LOOKUP_BY_POST_ID, TWITTER_BOOKMARKS_BY_USER, and TWITTER_GET_A_USER_S_OWNED_LISTS (as documented in SKILL.md) ingest content from the public internet into the agent's context.
Capability inventory: The skill possesses write and delete capabilities, including TWITTER_CREATION_OF_A_POST, TWITTER_POST_DELETE_BY_POST_ID, TWITTER_UNLIKE_POST, and TWITTER_REMOVE_A_BOOKMARKED_POST (as documented in SKILL.md).
Boundary markers: The instructions do not define specific delimiters or instructions for the agent to ignore potentially malicious content embedded within retrieved tweets.
Sanitization: There is no evidence of sanitization, filtering, or validation of the text retrieved from external sources before it is processed by the agent.
[SAFE]: The skill uses the Rube MCP server (https://rube.app/mcp) and Composio toolkit (composio.dev), which are established services for AI agent tool integration. The setup instructions for OAuth authentication through RUBE_MANAGE_CONNECTIONS follow standard security practices and avoid hardcoded credentials.

twitter-automation