Hi - I answer from the OpenSmartRoute documentation: routing, the API, plans and quotas, self-hosting. Ask away, or open a support ticket if you need a person.
Grounded in the docs - follow a source before acting on it.
build - Prompt - OpenSmartRoute
Promptv1.0.0
build
Build functionality by following plan and spec. Use caveman mode for all responses. Load relevant skills first, then implement incrementally with TDD.
Prompt file imported from ole-techwood/AGPWorkbook (.github/prompts/build.prompt.md). Copyright stays with the author.
Follow the caveman communication skill exactly as defined in SKILL.md. All responses must be terse, drop filler, use fragments. Active for every response in this session.
Load and follow SKILL.md first to choose relevant skills for task.
Do NOT use GPT-5 mini, GPT-5.4 mini, or any other "mini" OpenAI variant unless the user has explicitly requested one of those models by name or selected it via the model picker.
If Copilot is in Auto mode and classifies the task as simple, prefer Claude Haiku or Gemini Flash over any OpenAI mini model.
TASK
$input
Use it
Copy one of these into your project. Installing also returns the manifest and these snippets.
Follow the caveman communication skill exactly as defined in [SKILL.md](../../.agents/skills/caveman/SKILL.md). All responses must be terse, drop filler, use fragments. Active for every response in this session.
Load and follow [SKILL.md](../../.agents/skills/using-agent-skills/SKILL.md) first to choose relevant skills for task.
Required skills for this prompt:
- [SKILL.md](../../.agents/skills/clean-code/SKILL.md)
- [SKILL.md](../../.agents/skills/incremental-implementation/SKILL.md)
- [SKILL.md](../../.agents/skills/test-driven-development/SKILL.md)
- [SKILL.md](../../.agents/skills/context-engineering/SKILL.md)
- [SKILL.md](../../.agents/skills/doubt-driven-development/SKILL.md)
Optional skill when applicable for concrete task:
- [SKILL.md](../../.agents/skills/gof-patterns/SKILL.md)
Execution flow:
1. Read task acceptance criteria.
2. Load relevant context: existing code, patterns, types.
3. Write failing test for expected behavior (RED).
4. Implement minimum code to pass test (GREEN).
5. Run full test suite to check regressions.
6. Run build to verify compilation.
## Model Selection
**Do NOT use GPT-5 mini, GPT-5.4 mini, or any other "mini" OpenAI variant** unless the user has explicitly requested one of those models by name or selected it via the model picker.
If Copilot is in Auto mode and classifies the task as simple, prefer **Claude Haiku** or **Gemini Flash** over any OpenAI mini model.
## TASK
$input
Manifest
The prompt text and its fill-in variables. Copy it or fetch it by URL from your own code.
{
"prompt": "Follow the caveman communication skill exactly as defined in [SKILL.md](../../.agents/skills/caveman/SKILL.md). All responses must be terse, drop filler, use fragments. Active for every response in this session.\n\nLoad and follow [SKILL.md](../../.agents/skills/using-agent-skills/SKILL.md) first to choose relevant skills for task.\n\nRequired skills for this prompt:\n\n- [SKILL.md](../../.agents/skills/clean-code/SKILL.md)\n- [SKILL.md](../../.agents/skills/incremental-implementation/SKILL.md)\n- [SKILL.md](../../.agents/skills/test-driven-development/SKILL.md)\n- [SKILL.md](../../.agents/skills/context-engineering/SKILL.md)\n- [SKILL.md](../../.agents/skills/doubt-driven-development/SKILL.md)\n\nOptional skill when applicable for concrete task:\n\n- [SKILL.md](../../.agents/skills/gof-patterns/SKILL.md)\n\nExecution flow:\n\n1. Read task acceptance criteria.\n2. Load relevant context: existing code, patterns, types.\n3. Write failing test for expected behavior (RED).\n4. Implement minimum code to pass test (GREEN).\n5. Run full test suite to check regressions.\n6. Run build to verify compilation.\n\n## Model Selection\n\n**Do NOT use GPT-5 mini, GPT-5.4 mini, or any other \"mini\" OpenAI variant** unless the user has explicitly requested one of those models by name or selected it via the model picker.\n\nIf Copilot is in Auto mode and classifies the task as simple, prefer **Claude Haiku** or **Gemini Flash** over any OpenAI mini model.\n\n## TASK\n\n$input",
"variables": [],
"notes": "Build functionality by following plan and spec. Use caveman mode for all responses. Load relevant skills first, then implement incrementally with TDD.",
"metadata": {
"source": {
"provider": "github-prompts",
"repository": "https://github.com/ole-techwood/AGPWorkbook",
"path": ".github/prompts/build.prompt.md",
"ref": "49ffbf9259354b74ef121749f955b41f5135d23c",
"url": "https://github.com/ole-techwood/AGPWorkbook/blob/49ffbf9259354b74ef121749f955b41f5135d23c/.github/prompts/build.prompt.md",
"key": "ole-techwood/AGPWorkbook/.github/prompts/build.prompt.md"
}
}
}
Fetch it by URL: GET /api/v1/registry/ole-techwood-agpworkbook-build-prompt/manifest?version=1.0.0
Reviews
Star ratings from people who tried it. One review per account; edit yours any time.
No reviews yet. Install it, try it, and be the first to rate it.
