The Agent Skills Directory

[PROMPT_INJECTION]: The skill facilitates the ingestion of untrusted data from the Twitter API, creating a vulnerability to indirect prompt injection.
Ingestion points: User bios, tweet text, and search results are retrieved from external sources via user info, user tweets, and tweet search commands.
Boundary markers: Responses are structured as JSON, which provides a data container but lacks specific instructions to the agent to treat the content as data rather than instructions.
Capability inventory: The skill fetches content from the network and provides it directly to the agent's context for processing.
Sanitization: No sanitization, escaping, or filtering of the retrieved text content is documented or demonstrated.
[EXTERNAL_DOWNLOADS]: The skill interacts with the external service API at api.twitterapi.io and references documentation at docs.twitterapi.io.
[NO_CODE]: The primary logic for the tool is contained in scripts/twitterapi, which is described as a standalone Bun binary but was not included in the provided files for security verification.

twitterapi-cli