The Agent Skills Directory

[SAFE]: The skill provides structured instructions for product ideation and documentation. Analysis of the instructions shows no evidence of malicious behavior, prompt injection, or data exfiltration.
[EXTERNAL_DOWNLOADS]: The skill references two external URLs on GitHub belonging to the author's repository (github.com/MHammad33). These are provided as informational links for the user to visit after the ideation phase and are not downloaded or executed by the agent itself. Under internal safety rules, these references to the author's own documentation are considered safe.
[COMMAND_EXECUTION]: There are no shell commands or scripts included in this skill; it operates purely through natural language instructions for the LLM.
[DATA_EXFILTRATION]: No network operations or sensitive file accesses are performed. The skill only writes a file (IDEA.md) to the local working directory based on the conversation history.
[INDIRECT_PROMPT_INJECTION]: While the skill processes user-provided ideas, it does so within a structured framework and does not possess capabilities (like command execution or network access) that would allow for an exploitable injection surface.

idea