The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill directs users to install a CLI tool from a third-party GitHub repository and recommends adding more skills using npx skills add from unverified sources.
[REMOTE_CODE_EXECUTION]: The skill facilitates the execution of external code by recommending the installation of the infsh CLI and fetching remote skills via npx.
[COMMAND_EXECUTION]: The skill utilizes Bash to execute several shell commands, including infsh login, infsh app run, and npx skills add.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection (Category 8). Ingestion points: Processes external long-form content like blog posts and podcast transcripts (SKILL.md). Boundary markers: Absent; untrusted content is directly interpolated into tool arguments. Capability inventory: Executes shell commands via infsh app run to generate images, audio, and social media posts. Sanitization: Absent; no escaping or filtering is applied to external content before interpolation.

content-repurposing