The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process user-provided HTML files, which constitutes a potential attack surface for indirect prompt injection.
Ingestion points: The scripts scripts/seo_analyzer.py and scripts/generate_sitemap.py read the contents of HTML files within a user-specified directory.
Boundary markers: The scripts do not use specific delimiters or warnings to isolate processed content, instead providing structured reports that include data extracted from HTML tags (e.g., titles, meta descriptions, and headings).
Capability inventory: The skill performs file system reads and creates an XML sitemap file. It lacks network access and does not execute arbitrary shell commands.
Sanitization: The analysis logic uses a standard HTML parser to extract specific tag attributes but does not implement filtering for adversarial prompt instructions embedded in those tags.
[COMMAND_EXECUTION]: The skill instructions direct the agent to execute provided Python scripts for SEO auditing and sitemap creation. These scripts are local, deterministic, and serve the primary purpose of the skill without including dangerous execution patterns.
[SAFE]: No evidence of malicious behavior was detected. The skill does not hardcode credentials, utilize obfuscation techniques, attempt persistence, or download code from external sources.

seo-optimizer