The Agent Skills Directory

[REMOTE_CODE_EXECUTION] (CRITICAL): The skill contains the command curl -fsSL https://cli.inference.sh | sh. This pattern downloads a script from the internet and pipes it directly into a shell for execution without any verification. Because inference.sh is not an approved trusted source, this allows the remote server to execute arbitrary malicious code on the user's system.
[EXTERNAL_DOWNLOADS] (HIGH): The skill references several external packages using npx skills add inferencesh/skills@.... These are unverifiable dependencies from an untrusted third-party source, presenting a significant supply chain risk.
[PROMPT_INJECTION] (HIGH): This skill is vulnerable to Indirect Prompt Injection (Category 8). It is designed to ingest external content (blogs, podcasts) and repurpose it for social media.
Ingestion points: Processes external long-form source text via CLI tools.
Capability inventory: Has capabilities to post to social media (x/post-create) and run various generative AI models.
Boundary markers: None present in the provided examples.
Sanitization: No sanitization is mentioned. An attacker could embed instructions in a blog post that would cause the agent to post unauthorized content or exfiltrate information via the infsh tool.

content-repurposing