The Agent Skills Directory

[SAFE]: The skill implements strong security controls in its support scripts (e.g., append-loop-result.ts), including the use of realpathSync and explicit path prefix verification to prevent directory traversal attacks.
[SAFE]: No hardcoded credentials, API keys, or sensitive environment variable exposure patterns were found across the instructions or scripts.
[SAFE]: Command execution is restricted to local utility scripts shipped with the skill. Use of Bun.spawnSync is implemented with proper argument separation, mitigating risk of shell injection.
[SAFE]: The skill's external data ingestion (fetching robots.txt, sitemaps, and web pages for analysis) is handled by domain-specific agents whose outputs are verified by a sequential critic chain, reducing the impact of potential indirect prompt injections.
[SAFE]: Manifest and quality telemetry systems are implemented as local derived state, ensuring data persistence is scoped to the user's project directory.

optimize-seo