The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill directs the agent to ingest and analyze data from untrusted external sources, such as target JavaScript files and API responses, which creates a surface for indirect prompt injection attacks.
Ingestion points: Mapping URL parameters from targets and spidering external JavaScript source code.
Boundary markers: No specific delimiters are used to separate untrusted target data from agent instructions.
Capability inventory: The agent is tasked with executing shell commands (curl, ffuf) and network operations based on the results of target data analysis.
Sanitization: The methodology does not include specific validation or escaping steps for the content retrieved from target systems.
[COMMAND_EXECUTION]: The skill involves the use of command-line tools like curl, ffuf, and interactsh-client to perform security testing. While these are necessary for the skill's purpose, they provide the agent with powerful capabilities to interact with the network and execute local processes.

hunt-ssrf