The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill analyzes external websites and content, which introduces a potential surface for indirect prompt injection attacks where malicious instructions on a target page could influence the agent's behavior.
Ingestion points: Untrusted data enters the agent context through URL parameters in commands such as /aaron-marketing:audit [url] and /aaron-marketing:research [url] in SKILL.md.
Boundary markers: There are no explicit delimiters or warnings provided in the instructions to the agent to ignore instructions embedded within the processed content.
Capability inventory: The skill possesses capabilities for shell command execution (run.sh scripts) and network operations (via API connectors for SerpAPI, Ahrefs, etc.) as detailed in the Configuration and Advanced Usage sections of SKILL.md.
Sanitization: The documentation does not specify any sanitization or filtering of external content before it is processed by the marketing analysis skills.
[COMMAND_EXECUTION]: The skill orchestrates tasks by executing local shell scripts.
Evidence: The Advanced Usage section describes a bash wrapper that executes scripts located in the ./skills/ directory (e.g., ./skills/optimize/on-page-seo-auditor/run.sh).
[EXTERNAL_DOWNLOADS]: The installation instructions involve fetching code from external repositories and package registries.
Evidence: The skill recommends installation via npx skills add aaron-he-zhu/aaron-marketing-skills and git clone https://github.com/aaron-he-zhu/aaron-marketing-skills. These target a well-known service (GitHub) and common package registry tools.
[DYNAMIC_EXECUTION]: The skill dynamically invokes specific marketing sub-skills at runtime.
Evidence: Python examples demonstrate the use of a run_skill function to execute logic based on string identifiers (e.g., run_skill("keyword-research", ...)).

aaron-marketing-skills