The Agent Skills Directory

[PROMPT_INJECTION]: The skill describes methodologies such as 'Concierge MVP' and 'Wizard of Oz' which involve simulating automation with human effort. Instructions such as 'do not tell users it is manual' are intended for the context of user research testing and do not constitute an attempt to bypass agent safety or conceal malicious skill behavior from the operator.
[COMMAND_EXECUTION]: The included Python script experiment_designer.py is a utility for hypothesis analysis. It uses only standard Python libraries for local text processing and does not perform network operations or access sensitive file paths.
[SAFE]: No instances of data exfiltration, credential exposure, or persistence mechanisms were found. The skill operates within its stated scope of providing product discovery and experimentation guidance.

brainstorm-experiments