The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes external command-line utilities, specifically the bird CLI tool and smaug (via npx), to perform actions like fetching tweets and managing archives.
[EXTERNAL_DOWNLOADS]: Setup instructions require the installation of third-party software from external sources, including a Homebrew tap (steipete/tap/bird) and an NPM package (@steipete/bird), which are not from a verified trusted organization.
[DATA_EXFILTRATION]: The skill accesses sensitive local files, such as ~/.config/env/global.env and .env files, to retrieve X API credentials (Bearer tokens and OAuth keys) required for communication with the official X API.
[PROMPT_INJECTION]: The skill has a surface for indirect prompt injection because it processes and displays untrusted content (tweets) from external sources.
Ingestion points: Untrusted data is ingested via the X API and the bird scraping tool.
Boundary markers: There are no explicit delimiters or instructions provided to the agent to treat the fetched tweet content as data rather than instructions.
Capability inventory: The skill possesses capabilities for network communication, file system operations (reading credentials and writing research files), and subprocess execution.
Sanitization: The skill performs basic formatting and truncation of tweet text but lacks robust sanitization to prevent adversarial instructions within tweets from influencing the agent's behavior.

twitter