The Agent Skills Directory

[COMMAND_EXECUTION]: The skill orchestrates a development pipeline using the Bash tool to execute various commands, including npx agent-browser, vercel link, and vercel deploy. These operations are consistent with the skill's stated purpose of building and deploying websites and target well-known development services.- [PROMPT_INJECTION]: The skill contains an indirect prompt injection surface. It processes content extracted from external URLs (during Phase 1) and uses that data to inform the 'Plan' and 'Build' phases. If a target website contains malicious instructions embedded in its text or metadata, it could potentially influence the agent's code generation or deployment decisions.
Ingestion points: SKILL.md (Phase 2) reads structured data and content inventories extracted from external websites.
Boundary markers: Absent; there are no specific instructions for the agent to ignore or delimit embedded natural language instructions found in the source content.
Capability inventory: The agent has file-system access (writing code), shell execution (Bash), and deployment capabilities (Vercel).
Sanitization: No sanitization or validation of the extracted content is specified before it is used to generate the new site's structure and code.- [DATA_EXFILTRATION]: The skill uses vercel env pull to retrieve sensitive environment variables. This is a standard operation for the Vercel CLI used to sync local development with remote configuration, and within this context, it is considered legitimate functionality for a well-known service.

saccoai-website-rebuild