The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the user to install the belt CLI tool using npx skills add belt-sh/cli. It also references installation documentation hosted on GitHub at https://raw.githubusercontent.com/inference-sh/skills/refs/heads/main/cli-install.md.
[REMOTE_CODE_EXECUTION]: The core functionality of the skill is to send Python code for execution on the inference.sh cloud infrastructure. This behavior is transparently documented and serves as the primary feature of the skill.
[PROMPT_INJECTION]: The skill presents an attack surface for indirect prompt injection. Ingestion points: Python code in examples (e.g., in SKILL.md) retrieves content from external URLs using requests and BeautifulSoup. Boundary markers: The examples do not use delimiters or instructions to ignore embedded commands in scraped data. Capability inventory: The remote environment has access to a wide range of libraries for file manipulation and network access. Sanitization: No explicit sanitization or filtering of external content is demonstrated.
[SAFE]: The skill functions as a tool for developers to run scripts in a managed environment. No malicious patterns, obfuscation, or unauthorized access attempts were detected in the skill instructions or examples.

python-executor