The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill promotes an installation method using curl -fsSL https://cli.inference.sh | sh in SKILL.md, references/authentication.md, and references/cli-reference.md. This pattern downloads and executes a remote script directly in the shell without prior verification, posing a significant security risk.
[EXTERNAL_DOWNLOADS]: The skill fetches binaries and configuration files from cli.inference.sh and dist.inference.sh during the installation and update processes.
[DATA_EXFILTRATION]: The CLI automatically uploads local files to the inference.sh cloud when a file path is provided in the input JSON, which sends local data to third-party servers. This is documented as a core feature for AI processing in SKILL.md and references/running-apps.md.
[COMMAND_EXECUTION]: The skill is authorized to use the Bash tool to execute infsh commands, which perform network operations and handle local file uploads.
[CREDENTIALS_UNSAFE]: The documentation guides users to set an INFSH_API_KEY environment variable and explains that infsh login stores session credentials on the local filesystem, as described in references/authentication.md.
[PROMPT_INJECTION]: The skill processes untrusted data from third-party AI apps and local files, creating a surface for indirect prompt injection.
Ingestion points: App input fields and local file uploads (SKILL.md).
Boundary markers: Absent.
Capability inventory: Subprocess calls via infsh commands (SKILL.md).
Sanitization: Absent.

infsh-cli