The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill facilitates the installation of multiple third-party skill packs from community GitHub repositories, such as AgriciDaniel/claude-seo, addyosmani/web-quality-skills, aaron-he-zhu/seo-geo-claude-skills, coreyhaines31/marketingskills, jonathimer/devmarketing-skills, and blader/humanizer. It also uses the author's own French humanizer (samber/humaniseur-fr) and analytics proxy (hogpost.samber.dev).
[REMOTE_CODE_EXECUTION]: Employs the npx skills add command to fetch and integrate executable logic from external sources into the agent's environment. Although the skill mandates user confirmation before installation, the mechanism allows the execution of unverified community code.
[COMMAND_EXECUTION]: Executes various shell tools including dig, curl, openssl, and npx for infrastructure and security audits. It also dynamically generates a sub-agent configuration file and writes it to .claude/agents/weekly-seo.md.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection by ingesting untrusted data from remote URLs (site HTML, robots.txt, sitemaps) that could contain malicious instructions.
Ingestion points: Content from user-specified site URLs fetched via curl (SKILL.md, Phases 5, 6, and 7).
Boundary markers: Absent; external data is not delimited or accompanied by instructions to ignore embedded commands.
Capability inventory: The skill has access to Write, Edit, and npx skills add (SKILL.md).
Sanitization: Absent; external content is processed directly to verify site configuration and SEO metadata.

site-launch-checklist