The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill provides instructions for installing the CLI tool by fetching a shell script from the vendor's domain https://cli.inference.sh and piping it directly to the shell (sh).
[COMMAND_EXECUTION]: The skill configuration allows the agent to execute any command starting with infsh via the Bash tool to interact with the platform.
[DATA_EXFILTRATION]: The CLI tool includes functionality to read local files (e.g., images, video, audio) and upload them to the inference.sh servers for processing by AI applications.
[CREDENTIALS_UNSAFE]: The infsh login command facilitates authenticating with the service and stores the resulting credentials locally on the user's filesystem for persistent access.
[EXTERNAL_DOWNLOADS]: The manual installation instructions involve downloading binaries and metadata from https://dist.inference.sh/.
[PROMPT_INJECTION]: The skill presents an indirect prompt injection surface by ingesting and processing data from external AI applications. * Ingestion points: Untrusted data enters the context from the output of AI apps (e.g., web search results, LLM responses) via the infsh app run command. * Boundary markers: No explicit delimiters or instructions to ignore embedded commands are used when presenting external app output to the agent. * Capability inventory: The agent has access to the infsh tool which can perform network operations and access the local filesystem. * Sanitization: The skill does not implement sanitization or validation of the data retrieved from the remote AI services.

infsh-cli