The Agent Skills Directory

[COMMAND_EXECUTION]: The skill relies on shell commands to interact with the Twitter service through twitter-cli. These include post, search, me, and tweet operations.
[EXTERNAL_DOWNLOADS]: The skill documentation instructs the agent or user to install the twitter-cli tool using the uv package manager. This involves downloading and installing third-party software from a public registry.
[PROMPT_INJECTION]: The skill is vulnerable to Indirect Prompt Injection (Category 8) due to the following factors:
Ingestion points: The skill retrieves untrusted data from Twitter/X using the twitter-cli search and twitter-cli tweet <id> commands in SKILL.md.
Boundary markers: No specific boundary markers or instructions to ignore embedded directives are present in the prompt instructions.
Capability inventory: The agent has the capability to write back to the platform via twitter-cli post, which could be abused if an ingested tweet contains malicious instructions (SKILL.md).
Sanitization: There is no evidence of sanitization or filtering applied to the external data fetched from the social media platform before it is processed by the agent.

twitter-workflow