The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the agent to use npx -y pigeonscale, which results in downloading and executing the pigeonscale package from the npm registry at runtime.
[COMMAND_EXECUTION]: The skill is centered around executing shell commands using the pigeonscale CLI to perform operations such as site publishing, domain management, and configuration synchronization.
[DATA_EXFILTRATION]: The core purpose of the skill is to upload local directory contents to an external hosting provider (Pigeonscale). While this is the intended functionality, it constitutes the transfer of local data to a third-party service. The skill documentation mentions safety defaults to exclude sensitive files like .env, but also notes that these can be bypassed using the --skip-filter flag.
[INDIRECT_PROMPT_INJECTION]: This skill possesses an attack surface for indirect prompt injection as it processes and publicly hosts content from local directories which may be influenced by untrusted data.
Ingestion points: The skill reads files from user-specified local directories (e.g., ./dist) to be published.
Boundary markers: No explicit boundary markers or 'ignore' instructions are provided for the content being published.
Capability inventory: The skill executes shell commands and performs network uploads to pigeonscale.com.
Sanitization: The CLI tool includes default filtering for known sensitive file types (like .env) and performs a local secret scan, though these do not sanitize the content against prompt injection instructions.

instant-web-publishing-with-pigeonscale