Blackbox AI Prompts

Ask Blackbox AI to generate a [LANGUAGE] function that [FUNCTION_PURPOSE]. Inputs: [PARAM_1] ([TYPE_1]), [PARAM_2] ([TYPE_2]). It must handle the edge case [EDGE_CASE], return [RETURN_TYPE], and follow the [STYLE_GUIDE] style used in [EXISTING_FILE]. Have Blackbox explain the algorithm in [NUMBER] sentences, then write a quick test with [TEST_INPUT] to verify the output [EXPECTED_OUTPUT]. Review the generated code before using it in [PROJECT_NAME].

Extract the working code from this tutorial video using Blackbox AI: [VIDEO_URL_OR_FILE], focusing on the segment from [START_TIME] to [END_TIME]. The tutorial builds [FEATURE] in [LANGUAGE] with [FRAMEWORK]. Have Blackbox transcribe the code shown on screen, reconstruct it as a runnable [FILE_TYPE] file, and flag any steps the video skips (imports, config, environment variables). Test the extracted code in [ENVIRONMENT] and note what you had to fix.

Upload this design mockup [IMAGE_FILE] to Blackbox AI and ask it to generate the frontend code in [FRAMEWORK] with [STYLING, e.g. Tailwind CSS]. Requirements: responsive layout at [BREAKPOINT_1] and [BREAKPOINT_2], the [COMPONENT_NAME] component with [PROPS], and colors [COLOR_1] and [COLOR_2] matching the design tokens. Have it produce clean, commented code, then compare the render against the mockup and iterate on the [SECTION] that differs most.

Plan how to use Blackbox AI multi-model access (Claude, GPT, and Gemini variants) for my [PROJECT_TYPE] project. Assign Claude-class models to [TASK_1, e.g. architecture reviews], GPT-class models to [TASK_2, e.g. quick completions], and Gemini-class models to [TASK_3, e.g. long-context refactors]. Define when to switch models mid-task, how to compare outputs for [CRITICAL_FUNCTION], and a monthly credit budget of [BUDGET] on the [PLAN_NAME] plan so heavy use does not exhaust the allocation.

Run a pair-programming session with the Blackbox AI voice agent on [TASK_DESCRIPTION] in [LANGUAGE]. Speak the goal aloud, let the agent draft [FUNCTION_NAME], and review its suggestion on screen before accepting. Use voice commands to ask it to explain [COMPLEX_PART] and to refactor [CODE_SMELL]. Keep the session to [MINUTES] minutes, then summarize what the agent produced versus what you wrote manually in [NOTES_FILE].

Use Blackbox AI code search to find production-quality examples of [PATTERN_OR_ALGORITHM] in [LANGUAGE] across open-source repositories. Filter for examples with [CRITERIA, e.g. tests included, recent commits], compare [NUMBER] implementations, and pick the one that best fits [PROJECT_CONSTRAINT]. Adapt the chosen example to my codebase in [FILE_PATH], preserving its license attribution in [LICENSE_FILE]. Verify the adapted code passes [TEST_SUITE].

Paste this failing [LANGUAGE] code into Blackbox AI chat and ask it to find the bug: [CODE_SNIPPET]. The symptom is [SYMPTOM] when [TRIGGER]. Ask Blackbox to explain the root cause in plain language, propose [NUMBER] fixes ranked by safety, and show the corrected version of [FUNCTION_NAME]. Apply the safest fix to [FILE_PATH], re-run [TEST_COMMAND], and confirm the symptom is gone before closing the issue [ISSUE_ID].

Ask Blackbox AI to refactor the legacy [LANGUAGE] module [MODULE_NAME] (written in [OLD_STYLE]) to modern [NEW_STYLE] without changing behavior. The module handles [RESPONSIBILITY] and is covered by [TEST_SUITE]. Have Blackbox rewrite it file by file, run the tests after each file, and produce a summary of every API change in [CHANGELOG]. Flag any function whose behavior cannot be preserved exactly, and get my approval before it rewrites those.

Use Blackbox AI Pro Plus multi-agent execution to build [FEATURE_NAME] for [PROJECT_NAME]. Break the work into specialist steps: Agent 1 designs the [DATA_MODEL], Agent 2 implements [BACKEND_PART], Agent 3 builds [FRONTEND_PART], and Agent 4 writes [TESTS]. Give each agent the shared context file [CONTEXT_FILE], review each agent output before the next starts, and assemble the final result in [BRANCH_NAME]. Track credit usage against the [PLAN_NAME] plan budget.

Set up Blackbox AI across my development environments: [IDE_1], [IDE_2], and the terminal. Configure real-time autocomplete in [PRIMARY_IDE] with [LANGUAGE] as the priority language, enable code chat with my project at [REPO_PATH] as context, and set the browser extension to assist on [SITE_TYPE]. Define keyboard shortcuts for accept ([KEY_1]) and chat ([KEY_2]), and document the setup in [SETUP_DOC] so the team can replicate it.