The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes multiple local scripts including scripts/eval-runner.py, scripts/eval-config-manager.py, and scripts/quality-tracker.py. Specifically, in step 3 of the process, user-provided content is interpolated into a shell command: scripts/eval-runner.py --brand {slug} --action run-full --text "{content}". If the {content} variable contains shell metacharacters such as backticks or semicolons, it could result in arbitrary command execution.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to process and analyze untrusted content, which may contain malicious instructions intended to manipulate the evaluation results.
Ingestion points: The skill accepts user-provided text, files, or pasted blocks as 'Content to evaluate' in SKILL.md.
Boundary markers: Absent. The untrusted content is interpolated directly into command arguments without delimiters or 'ignore' instructions.
Capability inventory: The skill has the capability to run several Python scripts that perform complex logic, file reads from the home directory (~/.claude-marketing/), and write operations for logging (quality-tracker.py).
Sanitization: No evidence of sanitization, escaping, or validation of the {content} variable before it is passed to the execution environment.

eval-content