The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted text data to generate speech, which could be used as a vector for indirect prompt injection.
Ingestion points: The skill accepts user-provided text through the text parameter in the belt app run elevenlabs/tts command as seen in SKILL.md.
Boundary markers: The provided examples do not use delimiters or instructions to the agent to treat the input as pure data rather than potential instructions.
Capability inventory: The skill utilizes the Bash tool to execute the belt CLI tool, which interfaces with external APIs.
Sanitization: No input sanitization or filtering logic is present in the skill documentation to prevent potentially malicious text from being processed.
[EXTERNAL_DOWNLOADS]: The skill refers to external dependencies and installation instructions from third-party sources.
The documentation points to a remote installation script/guide located at https://raw.githubusercontent.com/inference-sh/skills/refs/heads/main/cli-install.md.
It recommends the installation of the belt-sh/cli and related inference-sh/skills using the npx package runner.

elevenlabs-tts