The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a local Python script scripts/quality-tracker.py using subprocess calls. It interpolates variables such as {slug}, {period}, and {content_type} directly into the shell command (e.g., scripts/quality-tracker.py --brand {slug} --action get-trends --days {period}). This pattern creates a potential vulnerability where an attacker could inject shell metacharacters into these parameters to execute unauthorized commands if the inputs are not properly sanitized.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes untrusted historical evaluation data to generate strategic insights and recommendations.
Ingestion points: The skill reads evaluation data logs and brand-specific JSON profiles from ~/.claude-marketing/.
Boundary markers: No specific delimiters or instructions to ignore embedded commands within the evaluation data were identified.
Capability inventory: The skill possesses the ability to execute local scripts and access the file system.
Sanitization: There is no evidence of sanitization or filtering of the content within the evaluation logs before they are synthesized into the final report.

quality-report