The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to use scripts/extract_product.py to fetch content from external e-commerce URLs provided by users. This creates a vulnerability surface where an attacker-controlled webpage could include hidden instructions designed to manipulate the agent's behavior during the content generation phase.
Ingestion points: External HTML content fetched via requests in scripts/extract_product.py.
Boundary markers: None identified; the agent is not instructed to treat the extracted product data as untrusted or to ignore embedded instructions.
Capability inventory: The skill includes network access via the extraction script and the ability to generate marketing text.
Sanitization: The script implements length limits on extracted text (e.g., text[:500]) but lacks sanitization or filtering to detect or remove potential prompt injection payloads.
[COMMAND_EXECUTION]: The skill instructions specify the execution of a local Python script scripts/extract_product.py to process product information. While this is a core part of the skill's functionality, it involves the execution of code within the agent's environment.
[EXTERNAL_DOWNLOADS]: The scripts/extract_product.py file requires the installation of external Python libraries requests and beautifulsoup4 to function. These are well-known, industry-standard libraries for network operations and HTML parsing.

instagram-marketing