The Agent Skills Directory

[PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface due to its handling of untrusted external content.
Ingestion points: Untrusted data is accepted through the materialUrl, materialFile, and materialText parameters defined in references/form-schema.json.
Boundary markers: The skill lacks delimiters or specific instructions to the underlying model to ignore instructions embedded within the ingested materials.
Capability inventory: The skill is primarily a network wrapper. Its capabilities are restricted to performing HTTPS requests to the vendor's API endpoint (https://ai-skills.ai/api/execute). No local shell execution or file modification capabilities were detected in the provided scripts.
Sanitization: The scripts/run.py script does not perform validation or sanitization of user-provided content before transmitting it to the API.
[SAFE]: All network activity is directed to the vendor's domain (ai-skills.ai) for the stated purpose of generating marketing ideas, using secure protocols.
[SAFE]: The execution script scripts/run.py follows standard Python practices and includes SSL certificate verification, optionally leveraging the certifi package for enhanced security.

ai-marketing-ideas