The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides detailed instructions for executing the bird command-line utility. This includes parameters for accessing browser profile directories (--chrome-profile-dir) and configuration files (~/.config/bird/config.json5) to extract session cookies for authentication.
[EXTERNAL_DOWNLOADS]: The skill metadata and documentation specify the installation of external software packages from the npm registry (@steipete/bird) and Homebrew (steipete/tap/bird). These are used to provide the core functionality of the skill.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it is designed to ingest and process untrusted data from external sources (X/Twitter tweets, replies, and search results).
Ingestion points: Data enters the agent context through commands like bird read, bird thread, bird home, and bird search.
Boundary markers: There are no explicit boundary markers or instructions to the agent to ignore embedded commands within the fetched social media content.
Capability inventory: The skill possesses capabilities that could be abused if an injection is successful, such as posting content (bird tweet), replying to users (bird reply), and modifying the social graph (bird follow).
Sanitization: No sanitization or filtering of the fetched social media content is described in the skill instructions.

bird