The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill recommends installing the belt CLI via npx skills add belt-sh/cli and references installation instructions hosted on a remote GitHub repository (raw.githubusercontent.com/inference-sh/skills/refs/heads/main/cli-install.md).
[REMOTE_CODE_EXECUTION]: The skill uses npx skills add to incorporate external packages such as inference-sh/skills@web-search and inference-sh/skills@prompt-engineering. This allows code from external sources to be added to the agent's environment.
[COMMAND_EXECUTION]: The skill utilizes the Bash tool to execute belt CLI commands for authentication (belt login) and running specialized research applications.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection as it retrieves data from external search providers (Tavily and Exa) to assist with fact-checking. Malicious content within the search results could potentially override agent instructions.
Ingestion points: Output from research tools including tavily/search-assistant, exa/search, and exa/answer (SKILL.md).
Boundary markers: None identified; the skill does not instruct the agent to ignore or delimit instructions found within the search results.
Capability inventory: The skill has permission to execute shell commands via the belt tool.
Sanitization: No sanitization or validation of the external search results is performed before processing.

press-release-writing