The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external Amazon product pages, which are processed by the agent to generate or optimize listings. This creates a surface where an attacker could embed malicious instructions in their product description to influence the agent's behavior.\n
Ingestion points: Product data retrieved via the fetch-listing.sh script and web_fetch from Amazon URLs specified in SKILL.md and README_CN.md.\n
Boundary markers: The skill does not implement explicit boundary markers or instructions for the agent to ignore potentially malicious commands within the fetched listing text.\n
Capability inventory: The skill possesses the ability to execute shell scripts and perform network operations via curl.\n
Sanitization: Although the fetch-listing.sh script uses sed to strip HTML tags, it does not sanitize the text for natural language instructions or prompt injection vectors.\n- [COMMAND_EXECUTION]: The skill relies on a local bash script located at scripts/fetch-listing.sh to retrieve data. This script is invoked with parameters derived from user input (ASIN and marketplace).\n- [EXTERNAL_DOWNLOADS]: The fetch-listing.sh script uses curl to retrieve content from Amazon's official regional domains. These requests are directed at well-known services necessary for the skill's function and are documented neutrally.

amazon-listing-optimization