The Agent Skills Directory

[DYNAMIC_CONTEXT_INJECTION]: The skill uses the !command syntax in SKILL.md to execute shell commands when the skill is loaded. These commands (command -v sqlmap, sqlmap --version, and test -f) are used to verify the presence of the tool and local wrapper scripts to inform the agent of the execution environment. These are benign status checks.
[COMMAND_EXECUTION]: The skill is designed to execute the sqlmap penetration testing tool and several local bash scripts (e.g., dump-database.sh, test-all-parameters.sh) located within the scripts/sqlmap/ directory.
[INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection as it processes content from external target URLs.
Ingestion points: Remote content retrieved from the target-url parameter during sqlmap execution.
Boundary markers: None explicitly defined in the skill instructions.
Capability inventory: Execution of command-line security tools and local shell scripts.
Sanitization: The skill mentions that commands are validated against a local .pentest/scope.json file using a PreToolUse hook.

sqlmap