The Agent Skills Directory

[SAFE]: The skill is well-structured and focuses on providing frameworks for adversarial thinking and decision-making. No malicious patterns, obfuscation, or unauthorized data exfiltration were detected.
[COMMAND_EXECUTION]: The skill utilizes local Python scripts (scripts/leadership_assessment.py, scripts/coaching_plan_generator.py, and scripts/goal_tracker.py) to process leadership scores and manage development goals. These scripts use standard libraries (argparse, json, datetime) and handle input safely without risky system calls or dynamic code evaluation.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user data (such as executive plans and business metrics) for analysis, which is an inherent risk for AI agents.
Ingestion points: User-provided business plans, strategic bets, and board meeting data.
Boundary markers: Absent; user plans are ingested directly into analysis frameworks without specific delimiters.
Capability inventory: The agent can execute local Python scripts and perform file writes to a local JSON database (goals_data.json) for persistence.
Sanitization: No explicit input validation or escaping mechanisms are implemented in the provided Python tools, though the risk is mitigated by the tools' limited scope.

executive-mentor