The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands using the sqlmap utility. It provides specific command patterns for GET and POST parameter testing, database enumeration, and data dumping.
[INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface where untrusted data from external sources is ingested and processed, specifically through the use of HTTP request files.
Ingestion points: The skill explicitly instructs the agent to use the sqlmap -r request.txt command, which reads the contents of an external file into the execution context.
Boundary markers: While the skill emphasizes manual triage and authorized scope, it lacks technical boundary markers or specific instructions to the agent to ignore embedded instructions within the ingested request file.
Capability inventory: The skill allows for significant system interaction, including executing complex database queries, enumerating schemas, and dumping table data.
Sanitization: There is no evidence of sanitization, validation, or escaping of the content within the request.txt file before it is passed to the tool.

sqlmap