The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to Indirect Prompt Injection because it processes content from external, untrusted websites.
Ingestion points: The scripts/seo_audit.py script fetches HTML content from arbitrary URLs. Additionally, various scripts fetch data from the DataForSEO API.
Boundary markers: The output from the audit and research scripts is presented to the agent without explicit boundary markers or instructions to ignore embedded commands.
Capability inventory: The skill possesses the capability to execute shell commands (curl), run local Python scripts, and perform network operations.
Sanitization: While scripts/seo_audit.py uses regex to extract specific tags, the extracted text (like titles and descriptions) is not sanitized for malicious instructions.
[COMMAND_EXECUTION]: The SKILL.md workflow instructs the agent to execute shell commands using curl and grep to inspect live websites, as well as several local Python scripts located in the scripts/ directory.
[EXTERNAL_DOWNLOADS]: The skill's primary function involves downloading content from external URLs (via curl and urllib.request) and interacting with the DataForSEO API endpoint (https://api.dataforseo.com/v3).
[REMOTE_CODE_EXECUTION]: A documentation file (references/schema-templates.md) contains a high-risk command pattern (curl -fsSL example.com/install.sh | bash) within a JSON-LD example. Although intended as a template for website FAQ content, it could be misinterpreted or accidentally executed by an AI agent.
[CREDENTIALS_UNSAFE]: The skill requires DATAFORSEO_LOGIN and DATAFORSEO_PASSWORD environment variables. These are handled using standard os.environ.get patterns in scripts/credential.py, which is generally acceptable but requires users to manage secrets locally.

seo-geo