Daniel Rosehill
@daniel-rosehill
Generative Ai
Negative prompt requesting that the agent refrain from creating unnecessary scripts and docs
Limits agent's role to generating repository documentation
Rule that defines scope of agent as documenting the repository (exact task depends on user prompt)
An artist trapped in a code generation bot ready to take your website design up a notch (or two)
Unpredictable and erratic - everything you didn't ask for you in your code generation assistant
William Shakespeare takes to the IDE as your coding assistant
Foundational role definition for my AI Assistants (assistant given a name for novelty; provide my name for personalisation)
Provides assistant with foundational environment info and tries to mitigate against tendency of AI agents to assume they're on a server they're not...
Uses the repository as teaching material to instruct the user in programming
Cybersecurity reviewer who comes with a grizzly past
General purpose config + personality direction for a code generation assistant that encourages users to take a sustainable approach to development work!
Evaluates the user's current tech stack and/or suggests alternatives or replacments
Instructs assistant to prepare repo for deployment to Streamlit Community Cloud
Instructs agent to use Docker on project
Assumes deployment to Cloudflare Pages
Block informing assistant that it will be working on feature additions which will be prompted
Instructs the assistant to honor a general preference for Postgres when architecting stacks
Conducts basic SEO review of repository
States that deployment will be onto Hugging Face
Instructs agent to use Streamlit
Instructs assistant to use Gradio
Instruction to use MongoDB as the database in the stack
Provide a list of general stack preferences to guide architecture/planning configs
Provides user's font preferences (for CSS/design editing configs)
Negative prompt asking agent to refrain from providing unsolicited security advice
Behavioral instruction asking agent to keep commentary to a minimum
Tells agent to use prompt.md as development prompt
Defines a list of user-preferred LLM APIs for projects
Sets deployment context with agent: this is not the server; it's a deployment repo
Writes a README page (no more, no less)
Ask the assistant to make sure that it's using the correct syntax for APIs and SDKs