The Agent Skills Directory

[REMOTE_CODE_EXECUTION] (CRITICAL): The skill explicitly instructs the agent or user to execute a remote script via a piped shell command (curl | sh). This allows the remote server to execute arbitrary code on the local system.
Evidence: SKILL.md contains curl -fsSL https://cli.inference.sh | sh.
Verification: This matches a confirmed detection pattern for untrusted remote code execution.
[COMMAND_EXECUTION] (HIGH): The skill relies on executing the infsh binary, which is a third-party executable downloaded at runtime without verification from a trusted platform.
[EXTERNAL_DOWNLOADS] (MEDIUM): The skill references and downloads assets from inference.sh and dist.inference.sh, which are not recognized as trusted sources.
[PROMPT_INJECTION] (LOW): The skill possesses an indirect prompt injection surface by taking untrusted data (URLs and prompts) and passing them as arguments to a CLI tool.
Ingestion points: image_url and prompt fields in JSON payloads.
Boundary markers: None present to distinguish between instructions and data.
Capability inventory: Command execution via Bash tool.
Sanitization: No sanitization logic is evident in the skill instructions.

image-upscaling