The Agent Skills Directory

[PRIVILEGE_ESCALATION]: The skill explicitly instructs the agent to bypass sandbox restrictions by requesting escalated permissions (sandbox_permissions=require_escalated) if network calls are blocked. While this facilitates deployment to a well-known service, it encourages bypassing security boundaries.
[METADATA_POISONING]: The skill's metadata contains conflicting authorship information. The YAML frontmatter claims the author is github.com/openai/skills, whereas the provided context indicates the author is tech-leads-club. This discrepancy could mislead users regarding the skill's origin and safety.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted project data that could contain malicious instructions.
Ingestion points: Reads package.json, netlify.toml, and Git remote configuration from the local project.
Boundary markers: The skill lacks explicit instructions or delimiters to ignore potential malicious content within these files.
Capability inventory: Includes execution of npx netlify and npm install, which could be exploited if malicious configuration is ingested.
Sanitization: No evidence of input validation or sanitization for project-controlled files.
[COMMAND_EXECUTION]: The skill relies on executing external commands such as npx netlify and npm install. While necessary for deployment, running npm install on potentially untrusted project directories poses a risk if malicious dependencies are present.
[EXTERNAL_DOWNLOADS]: The skill uses npx netlify to download and execute the Netlify CLI package from the npm registry. While Netlify is a well-known service, this involves runtime code acquisition.

netlify-deploy