The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes powerful system tools, including Bash, Write, and Edit. Its core philosophy ('Speed Over Perfection', 'Bias Toward Action') explicitly encourages the agent to bypass careful analysis and 'just get it working.' This creates a risk where the agent might execute potentially dangerous shell commands or modify critical files in an attempt to quickly resolve a 'blocker' without fully understanding the security implications.
[PROMPT_INJECTION]: The skill's instructions create a vulnerability surface for indirect prompt injection (Category 8). An attacker could provide a task that appears to be a simple 'blocker' but contains hidden instructions that the agent might execute without deliberation due to its '15-Minute Rule' and mission to be the 'antidote to analysis paralysis.'
Ingestion points: External data and task descriptions processed through Read, Grep, and Glob tools.
Boundary markers: No specific delimiters or safety instructions are provided to distinguish between the skill's system instructions and untrusted user data.
Capability inventory: Full access to Bash command execution, file system writing, and editing.
Sanitization: The workflow lacks any sanitization or validation steps, explicitly favoring immediate implementation ('Stop thinking, start doing') over security review.

swift-executor