The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill requires the installation of a third-party Python package blockrun-llm from a public registry. This dependency is not from a well-known service or trusted organization.
[COMMAND_EXECUTION]: The skill utilizes Bash tools to install the required SDK, execute Python scripts for model interaction, and manage wallet session files in the user's home directory.
[INDIRECT_PROMPT_INJECTION]: The skill acts as a bridge for user-supplied data to external LLMs, creating a surface for potential prompt injection attacks.
Ingestion points: User prompts are directly passed to client.chat() and client.generate() functions within the Python code examples in SKILL.md.
Boundary markers: The instructions do not define delimiters or warnings to prevent the downstream models from following malicious instructions embedded in the user data.
Capability inventory: The skill has permissions for file reading, shell command execution (Python/Pip), and network connectivity via the installed SDK.
Sanitization: There is no evidence of input validation, filtering, or sanitization before external transmission.

blockrun