The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill provides an interface to ingest data from an external source (Pepper Cloud), which represents a potential attack surface for indirect prompt injection.
Ingestion points: Data is retrieved from the $PEPPER_CLOUD_URL/api/state endpoint via the state_read function in SKILL.md.
Boundary markers: The skill does not define specific delimiters (e.g., XML tags) to wrap retrieved content when it is loaded into the agent's context (e.g., in the COMPANY or STRATEGY variables).
Capability inventory: The skill uses curl for network requests and python3 for data processing. It does not perform high-risk actions like eval() or arbitrary file system writes directly on the ingested content within this file.
Sanitization: The skill correctly utilizes urllib.parse.quote for URL parameter safety and json.dumps/json.load for safe JSON handling, preventing basic command injection in the transport layer.

pepper-state