The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes external market data fetched via search tools, creating a surface for instructions embedded in search results to influence agent behavior.
Ingestion points: Results from tavily/search-assistant and exa/search tools are incorporated into the agent context in SKILL.md.
Boundary markers: Absent; the skill does not use delimiters or explicit instructions to ignore potentially malicious prompts within the researched data.
Capability inventory: The skill maintains access to the Bash tool (restricted to belt commands), enabling network interactions and application execution.
Sanitization: No evidence of validation or filtering for the data retrieved from external search providers.
[EXTERNAL_DOWNLOADS]: The skill references an external installation script and configuration guide located at https://raw.githubusercontent.com/inference-sh/skills/refs/heads/main/cli-install.md.
[REMOTE_CODE_EXECUTION]: The instructions direct the user to install external packages and tools from non-standard sources, including belt-sh/cli and various inference-sh skills via npx skills add.
[COMMAND_EXECUTION]: The skill utilizes the Bash tool to execute belt CLI commands, which facilitate communication with the inference.sh platform and run remote applications.

customer-persona