:hedgehog: PostHog is the leading platform for building self-driving products. Our developer tools – AI observability, analytics, session replay, flags, experiments, error tracking, logs, and more – capture all the context agents need to diagnose problems, uncover opportunities, and ship fixes. Steer it all from Slack, web, desktop, or the MCP.
Python ab-testingai-analyticsanalyticscdpdata-warehouseexperimentsfeature-flagsjavascriptproduct-analyticspythonreactsession-replaysurveystypescriptweb-analytics
import { describe, expect, it, vi } from 'vitest'
import { getToolByName } from '@/shared/test-utils'
import { GENERATED_TOOLS } from '@/tools/generated/ai_observability'
import type { Context } from '@/tools/types'
function createContext(requestReturnValue: unknown): { context: Context; requestMock: ReturnType<typeof vi.fn> } {
const requestMock = vi.fn().mockResolvedValue(requestReturnValue)
const context = {
api: {
request: requestMock,
},
stateManager: {
getProjectId: vi.fn().mockResolvedValue('17'),
},
} as unknown as Context
return { context, requestMock }
}
describe('Generated llma-prompt-* tools', () => {
// prompt-get also accepts `name` and the other identifier aliases, normalized to `prompt_name`
// before the handler sees them — see llma-identifier-contracts.test.ts.
it('uses prompt_name in generated prompt-get schema', () => {
const tool = getToolByName(GENERATED_TOOLS, 'llma-prompt-get')
const parsed = tool.schema.parse({ prompt_name: 'checkout_prompt', version: 2 })
expect(parsed).toEqual({ prompt_name: 'checkout_prompt', version: 2, content: 'full', resolve: true })
})
it('exposes content mode on the prompt-get schema so agents can fetch outline-only', () => {
const tool = getToolByName(GENERATED_TOOLS, 'llma-prompt-get')
const parsed = tool.schema.parse({ prompt_name: 'checkout_prompt', content: 'none' })
expect(parsed).toEqual({ prompt_name: 'checkout_prompt', content: 'none', resolve: true })
expect(() => tool.schema.parse({ prompt_name: 'checkout_prompt', content: 'bogus' })).toThrow()
})
it('uses prompt_name (not name) in generated prompt-update schema', () => {
const tool = getToolByName(GENERATED_TOOLS, 'llma-prompt-update')
const parsed = tool.schema.parse({ prompt_name: 'checkout_prompt', prompt: { text: 'v2' }, base_version: 1 })
expect(parsed).toEqual({ prompt_name: 'checkout_prompt', prompt: { text: 'v2' }, base_version: 1 })
expect(() => tool.schema.parse({ name: 'checkout_prompt', prompt: { text: 'v2' } })).toThrow()
})
it('requires base_version on the prompt-update schema (PATCH optionality stripped)', () => {
const tool = getToolByName(GENERATED_TOOLS, 'llma-prompt-update')
// The backend serializer requires base_version for optimistic concurrency, but the
// PATCH endpoint marks it optional in OpenAPI — the schema must re-require it so the
// agent can't silently omit it and hit a backend 400.
expect(() => tool.schema.parse({ prompt_name: 'checkout_prompt', prompt: { text: 'v2' } })).toThrow()
})
it('wires prompt-list to GET and preserves paginated output shape', async () => {
const paginated = {
count: 1,
next: null,
previous: null,
results: [
{
id: '2f53a52a-06f5-4025-9ea7-f763f74f17f5',
name: 'checkout_prompt',
version: 3,
prompt_size_bytes: 123,
},
],
}
const { context, requestMock } = createContext(paginated)
const tool = getToolByName(GENERATED_TOOLS, 'llma-prompt-list')
const result = await tool.handler(context, { search: 'checkout' })
expect(requestMock).toHaveBeenCalledWith({
method: 'GET',
path: '/api/projects/17/llm_prompts/',
query: { search: 'checkout', content: 'none' },
})
expect(result).toEqual(paginated)
})
it('allows overriding prompt-list content mode', async () => {
const { context, requestMock } = createContext({ count: 0, next: null, previous: null, results: [] })
const tool = getToolByName(GENERATED_TOOLS, 'llma-prompt-list')
await tool.handler(context, { search: 'checkout', content: 'preview' })
expect(requestMock).toHaveBeenCalledWith({
method: 'GET',
path: '/api/projects/17/llm_prompts/',
query: { search: 'checkout', content: 'preview' },
})
})
it('wires prompt-get to GET /name/{prompt_name}/ with version query', async () => {
const response = { id: '2f53a52a-06f5-4025-9ea7-f763f74f17f5', name: 'checkout_prompt', version: 2 }
const { context, requestMock } = createContext(response)
const tool = getToolByName(GENERATED_TOOLS, 'llma-prompt-get')
await tool.handler(context, { prompt_name: 'checkout_prompt', version: 2 })
expect(requestMock).toHaveBeenCalledWith({
method: 'GET',
path: '/api/projects/17/llm_prompts/name/checkout_prompt/',
query: { version: 2 },
})
})
it('wires prompt-create to POST body with name and prompt', async () => {
const { context, requestMock } = createContext({ id: 'new-id' })
const tool = getToolByName(GENERATED_TOOLS, 'llma-prompt-create')
await tool.handler(context, { name: 'new_prompt', prompt: { text: 'hello' } })
expect(requestMock).toHaveBeenCalledWith({
method: 'POST',
path: '/api/projects/17/llm_prompts/',
body: { name: 'new_prompt', prompt: { text: 'hello' } },
})
})
it('wires prompt-update to PATCH /name/{prompt_name}/ with publish payload', async () => {
const { context, requestMock } = createContext({ id: 'updated-id' })
const tool = getToolByName(GENERATED_TOOLS, 'llma-prompt-update')
await tool.handler(context, { prompt_name: 'new_prompt', prompt: { text: 'v2' }, base_version: 1 })
expect(requestMock).toHaveBeenCalledWith({
method: 'PATCH',
path: '/api/projects/17/llm_prompts/name/new_prompt/',
body: { prompt: { text: 'v2' }, base_version: 1 },
})
})
it('uses prompt_name and new_name in generated prompt-duplicate schema', () => {
const tool = getToolByName(GENERATED_TOOLS, 'llma-prompt-duplicate')
const parsed = tool.schema.parse({ prompt_name: 'original_prompt', new_name: 'copy_of_prompt' })
expect(parsed).toEqual({ prompt_name: 'original_prompt', new_name: 'copy_of_prompt' })
expect(() => tool.schema.parse({ prompt_name: 'original_prompt' })).toThrow()
})
it('wires prompt-duplicate to POST /name/{prompt_name}/duplicate/ with new_name body', async () => {
const response = { id: 'new-id', name: 'copy_of_prompt', version: 1 }
const { context, requestMock } = createContext(response)
const tool = getToolByName(GENERATED_TOOLS, 'llma-prompt-duplicate')
const result = await tool.handler(context, { prompt_name: 'original_prompt', new_name: 'copy_of_prompt' })
expect(requestMock).toHaveBeenCalledWith({
method: 'POST',
path: '/api/projects/17/llm_prompts/name/original_prompt/duplicate/',
body: { new_name: 'copy_of_prompt' },
})
expect(result).toEqual(response)
})
})