The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the AI agent to construct and execute shell commands using grep and curl. These commands are used to parse local log files and perform active network requests against external targets.
[EXTERNAL_DOWNLOADS]: The skill utilizes curl to perform network operations against external domains (represented by the placeholder target.com). This enables the agent to interact with non-whitelisted remote servers to send SQL injection payloads.
[PROMPT_INJECTION]: The skill is highly vulnerable to indirect prompt injection due to its core processing logic.
Ingestion points: The agent is instructed to read and analyze log.txt, which contains captured network traffic (untrusted data from external sources).
Boundary markers: There are no delimiters or instructions to ignore embedded commands or instructions within the captured traffic.
Capability inventory: The agent has access to grep for pattern matching and curl for making network requests across the instructions.
Sanitization: There is no evidence of sanitization, escaping, or validation of the content extracted from log.txt before it is interpolated into shell commands or used to drive subsequent agent actions.
[COMMAND_EXECUTION]: The use of $ARGUMENTS in the markdown header represents a potential command injection vector if the platform's execution environment does not properly sanitize template variables before they are processed by a shell.

mitm-find-sqli