The Agent Skills Directory

[SAFE]: The skill provides instructional guidance for a development workflow without including any executable code, network requests, or sensitive data access. The instructions emphasize safety by directing the agent to clarify vague or unsafe scopes before proceeding.
[PROMPT_INJECTION]: The skill defines agent behavior for software engineering tasks but does not contain patterns attempting to bypass safety filters, disregard prior instructions, or extract system prompts.
[DATA_EXFILTRATION]: No network operations, hardcoded credentials, or sensitive file path access were identified in the instructions.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to process external inputs such as issue tracker items (e.g., Linear) and task descriptions. While this creates a surface for indirect prompt injection, the instructions explicitly command the agent to clarify or break down scope if it is vague or unsafe, and to verify behavior through tests, mitigating the risk of accidental or malicious command execution from external data.

implement