The Agent Skills Directory

[SAFE]: The skill contains standard development instructions focused on code quality, runtime correctness, and testing. There are no patterns suggesting malicious intent, data exfiltration, or unauthorized command execution.
[NO_CODE]: This skill consists entirely of Markdown instructions and metadata. It does not include any accompanying scripts (Bash, Python, JS), configuration files, or executable binaries, which significantly reduces the attack surface.
[PROMPT_INJECTION]: While the skill uses instructional language like 'Prioritize' and 'Do not convert', these are legitimate task constraints rather than attempts to bypass agent safety filters or override system-level instructions.
[INDIRECT_PROMPT_INJECTION]: The skill presents an indirect prompt injection surface because it is designed to analyze and modify existing codebases via workspace-write capabilities.
Ingestion points: The agent is instructed to map execution boundaries by reading existing workspace files.
Boundary markers: There are no explicit instructions or delimiters defined to distinguish code from potential instructions embedded within comments or data strings in the files being analyzed.
Capability inventory: The skill utilizes workspace-write to implement fixes and recommendations.
Sanitization: No specific sanitization or filtering logic is provided to handle untrusted input from the codebase, though the risk is considered low and inherent to coding assistance tasks.

javascript-pro