The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill instructs users to install its core dependency by piping a remote script from https://cli.inference.sh directly into the shell (curl -fsSL ... | sh). This pattern is a high-risk security practice as it allows for the execution of arbitrary, unverified code on the host system.
[EXTERNAL_DOWNLOADS]: The skill downloads binary executables and additional configuration files from the inference.sh and dist.inference.sh domains during both installation and runtime operations.
[COMMAND_EXECUTION]: The skill relies on the execution of the infsh CLI tool to perform its primary functions. This involves spawning subprocesses and passing user-provided or externally sourced data as command-line arguments and JSON payloads.
[PROMPT_INJECTION]: The skill exhibits a significant surface for indirect prompt injection by design. It is built to ingest and process long-form external content such as blog posts, podcast transcripts, and video scripts from potentially untrusted sources.
Ingestion points: Workflow instructions in SKILL.md identify blog posts, podcasts, and video transcripts as primary data inputs.
Boundary markers: No specific delimiters or instructions to ignore embedded commands are present in the provided tool call examples.
Capability inventory: The skill uses the infsh tool which has capabilities including image generation, text-to-speech, and social media account interaction (e.g., posting to X/Twitter).
Sanitization: There is no evidence of input validation, escaping, or sanitization before external content is interpolated into the CLI command inputs.

content-repurposing