The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to process and analyze external auction notices, which are untrusted data sources that could potentially contain adversarial instructions.
Ingestion points: The agent is instructed to read auction notice documents and extract specific data points as described in the protocol within references/detailed-guide.md.
Boundary markers: There are no explicit instructions to use delimiters or "ignore" blocks to isolate the untrusted document content from the agent's core instructions during analysis.
Capability inventory: The skill is granted access to powerful tools in its configuration, such as claude-code and gemini-cli, which possess capabilities for file system interaction and command execution.
Sanitization: The protocol does not include steps to sanitize, escape, or validate the content of the auction notices before the agent processes them.

leiloeiro-edital