The Agent Skills Directory

[DATA_EXFILTRATION]: The skill is instructed to access and write to sensitive local file paths (e.g., ~/.config/readx/credentials.json or %APPDATA%\readx\credentials.json) to store and retrieve the READX_API_KEY.
[COMMAND_EXECUTION]: The toolkit utilizes Bash and curl to interact with the ReadX API when MCP tools are unavailable, which requires the agent to execute shell commands directly.
[EXTERNAL_DOWNLOADS]: The skill dynamically fetches documentation from a remote source (https://readx.cc/api-docs.txt) and uses the content to guide its API interactions, introducing a dependency on remote instruction integrity.
[PROMPT_INJECTION]: The skill's primary function involves analyzing untrusted external data from Twitter/X, creating a surface for indirect prompt injection attacks.
Ingestion points: Tweet content, user profiles, community metadata, and search results retrieved from the Twitter/X API.
Boundary markers: The instructions do not specify the use of delimiters or clear directives for the agent to ignore instructions embedded within the social media data.
Capability inventory: The agent has the ability to read/write local configuration files and execute shell commands (curl).
Sanitization: There are no explicit steps provided to sanitize or filter retrieved content before it is processed by the agent's logic.

readx