The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill's primary function involves ingesting and analyzing untrusted content from external URLs and Google Search results. This inherent surface area could allow external content to attempt to influence the agent's behavior during analysis. However, the risk is minimal as the skill lacks dangerous capabilities like arbitrary file writing or outbound network access beyond search APIs.
Ingestion points: SKILL.md Step 1 (Target URL content) and Step 2 (Google SERP snippets and People Also Ask questions).
Boundary markers: Not explicitly defined in the prompts.
Capability inventory: Limited to standard WebSearch, DataForSEO API calls, and local parsing scripts for text analysis.
Sanitization: Not specified for the external data.
[COMMAND_EXECUTION]: The skill uses local scripts (render_page.py and parse_html.py) to process user-provided URLs. The instruction format assumes the execution environment safely handles the URL argument to prevent shell-based command injection.
[SAFE]: The behaviors described in the instructions are entirely consistent with the skill's stated purpose as an SEO/SXO utility. The use of local scripts within the ${CLAUDE_PLUGIN_ROOT} directory and the logic for persona derivation and gap scoring follow standard industry practices without exhibiting malicious patterns.

seo-sxo