Merge remote-tracking branch 'origin/dev' into opencode/mighty-wolf

This commit is contained in:
Choi Kijin / 최 기진 / チョイ キジン
2026-04-28 15:47:58 +09:00
141 changed files with 6764 additions and 932 deletions
@@ -1,6 +1,7 @@
import { describe, it, expect } from "bun:test"
import {
BOULDER_CONTINUATION_PROMPT,
SINGLE_TASK_DIRECTIVE,
VERIFICATION_REMINDER,
VERIFICATION_REMINDER_GEMINI,
} from "./system-reminder-templates"
@@ -32,8 +33,8 @@ describe("BOULDER_CONTINUATION_PROMPT", () => {
expect(checkboxMarkingMatch).not.toBeNull()
expect(proceedMatch).not.toBeNull()
const checkboxPosition = checkboxMarkingMatch!.index
const proceedPosition = proceedMatch!.index
const checkboxPosition = checkboxMarkingMatch!.index ?? -1
const proceedPosition = proceedMatch!.index ?? -1
expect(checkboxPosition).toBeLessThan(proceedPosition)
})
@@ -51,3 +52,19 @@ describe("VERIFICATION_REMINDER_GEMINI", () => {
expect(VERIFICATION_REMINDER_GEMINI).toContain(":!node_modules")
})
})
describe("SINGLE_TASK_DIRECTIVE", () => {
it("does not contain refusal language", () => {
// given
const lowerCaseDirective = SINGLE_TASK_DIRECTIVE.toLowerCase()
// when / then
expect(lowerCaseDirective).not.toContain("refuse")
expect(SINGLE_TASK_DIRECTIVE).not.toContain("I refuse")
})
it("contains systematic execution guidance", () => {
expect(SINGLE_TASK_DIRECTIVE).toContain("EXECUTION PROTOCOL")
expect(SINGLE_TASK_DIRECTIVE).toContain("VERIFICATION IS MANDATORY")
})
})
+16 -23
View File
@@ -217,33 +217,26 @@ export const SINGLE_TASK_DIRECTIVE = `
${createSystemDirective(SystemDirectiveTypes.SINGLE_TASK_ONLY)}
**STOP. READ THIS BEFORE PROCEEDING.**
**EXECUTION PROTOCOL**
If you were given **multiple genuinely independent goals** (unrelated tasks, parallel workstreams, separate features), you MUST:
1. **IMMEDIATELY REFUSE** this request
2. **DEMAND** the orchestrator provide a single goal
Work systematically. Each unit must be verified before proceeding.
**What counts as multiple independent tasks (REFUSE):**
- "Implement feature A. Also, add feature B."
- "Fix bug X. Then refactor module Y. Also update the docs."
- Multiple unrelated changes bundled into one request
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
**What is a single task with sequential steps (PROCEED):**
- A single goal broken into numbered steps (e.g., "Implement X by: 1. finding files, 2. adding logic, 3. writing tests")
- Multi-step context where all steps serve ONE objective
- Orchestrator-provided context explaining approach for a single deliverable
| Step | Action | Verification |
|------|--------|--------------|
| 1 | Identify first atomic unit | Smallest complete piece of work |
| 2 | Execute fully | Implement the change |
| 3 | Verify | \`lsp_diagnostics\`, tests, build |
| 4 | Report | State what's done, what remains |
| 5 | Continue | Next unit, or await if scope unclear |
**Your response if genuinely independent tasks are detected:**
> "I refuse to proceed. You provided multiple independent tasks. Each task needs full attention.
>
> PROVIDE EXACTLY ONE GOAL. One deliverable. One clear outcome.
>
> Batching unrelated tasks causes: incomplete work, missed edge cases, broken tests, wasted context."
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
**WARNING TO ORCHESTRATOR:**
- Bundling unrelated tasks RUINS deliverables
- Each independent goal needs FULL attention and PROPER verification
- Batch delegation of separate concerns = sloppy work = rework = wasted tokens
**VERIFICATION IS MANDATORY.** No skipping. No batching completions.
**REFUSE genuinely multi-task requests. ALLOW single-goal multi-step workflows.**
**IF SCOPE SEEMS BROAD:**
Complete the first logical unit. Report progress. Await further instruction if needed.
**REMEMBER:** Prometheus already decomposed the work. Execute what you receive.
`
@@ -1,8 +1,9 @@
/// <reference types="bun-types" />
import { afterEach, describe, expect, it, spyOn } from "bun:test"
import type { LoadedSkill } from "../../features/opencode-skill-loader"
import * as shared from "../../shared"
import * as slashcommand from "../../tools/slashcommand"
import { executeSlashCommand } from "./executor"
import * as slashcommand from "../../tools/slashcommand/command-discovery"
let resolveCommandsInTextSpy: { mockRestore: () => void } | undefined
let resolveFileReferencesInTextSpy: { mockRestore: () => void } | undefined
@@ -38,6 +39,11 @@ function restoreExecutorSpies(): void {
discoverCommandsSyncSpy = undefined
}
async function executeSlashCommand(...args: Parameters<typeof import("./executor").executeSlashCommand>): ReturnType<typeof import("./executor").executeSlashCommand> {
const module = await import(`./executor?test=${Date.now()}-${Math.random()}`)
return module.executeSlashCommand(...args)
}
afterEach(restoreExecutorSpies)
function createRestrictedSkill(): LoadedSkill {
+4 -6
View File
@@ -1,10 +1,8 @@
import { dirname } from "path"
import {
resolveCommandsInText,
resolveFileReferencesInText,
} from "../../shared"
import { resolveCommandsInText } from "../../shared/command-executor/resolve-commands-in-text"
import { resolveFileReferencesInText } from "../../shared/file-reference-resolver"
import { discoverAllSkills, type LoadedSkill, type LazyContentLoader } from "../../features/opencode-skill-loader"
import { discoverCommandsSync } from "../../tools/slashcommand"
import * as commandDiscovery from "../../tools/slashcommand/command-discovery"
import type { CommandInfo as DiscoveredCommandInfo, CommandMetadata } from "../../tools/slashcommand/types"
import type { ParsedSlashCommand } from "./types"
@@ -47,7 +45,7 @@ export interface ExecutorOptions {
async function discoverAllCommands(options?: ExecutorOptions): Promise<CommandInfo[]> {
const discoveredCommands = discoverCommandsSync(options?.directory ?? process.cwd(), {
const discoveredCommands = commandDiscovery.discoverCommandsSync(options?.directory ?? process.cwd(), {
pluginsEnabled: options?.pluginsEnabled,
enabledPluginsOverride: options?.enabledPluginsOverride,
})
+31 -5
View File
@@ -1,18 +1,18 @@
import type { PluginInput } from "@opencode-ai/plugin"
import { isGptModel, isGpt5_4Model } from "../../agents/types"
import { isGptModel, isGptNativeSisyphusModel } from "../../agents/types"
import {
getSessionAgent,
resolveRegisteredAgentName,
updateSessionAgent,
} from "../../features/claude-code-session-state"
import { log } from "../../shared"
import { AGENT_MODEL_REQUIREMENTS, log } from "../../shared"
import { getAgentConfigKey } from "../../shared/agent-display-names"
const TOAST_TITLE = "NEVER Use Sisyphus with GPT"
const TOAST_MESSAGE = [
"Sisyphus works best with Claude Opus, and works fine with Kimi/GLM models.",
"Do NOT use Sisyphus with GPT (except GPT-5.4 which has specialized support).",
"For GPT models (other than 5.4), always use Hephaestus.",
"Do NOT use Sisyphus with GPT (except GPT-5.4 and GPT-5.5 which have specialized support).",
"For other GPT models, always use Hephaestus.",
].join("\n")
function showToast(ctx: PluginInput, sessionID: string): void {
ctx.client.tui.showToast({
@@ -30,6 +30,18 @@ function showToast(ctx: PluginInput, sessionID: string): void {
})
}
function getNativeSisyphusGptVariant(model: { providerID: string; modelID: string }): string | undefined {
const chain = AGENT_MODEL_REQUIREMENTS["sisyphus"]?.fallbackChain ?? []
const exactMatch = chain.find((entry) =>
entry.providers.includes(model.providerID) && entry.model === model.modelID
)
if (exactMatch?.variant !== undefined) {
return exactMatch.variant
}
return chain.find((entry) => entry.model === model.modelID)?.variant
}
export function createNoSisyphusGptHook(ctx: PluginInput) {
return {
"chat.message": async (input: {
@@ -43,7 +55,21 @@ export function createNoSisyphusGptHook(ctx: PluginInput) {
const agentKey = getAgentConfigKey(rawAgent)
const modelID = input.model?.modelID
if (agentKey === "sisyphus" && modelID && isGptModel(modelID) && !isGpt5_4Model(modelID)) {
if (
agentKey === "sisyphus"
&& input.model
&& modelID
&& isGptNativeSisyphusModel(modelID)
&& output?.message
&& output.message.variant === undefined
) {
const variant = getNativeSisyphusGptVariant(input.model)
if (variant !== undefined) {
output.message.variant = variant
}
}
if (agentKey === "sisyphus" && modelID && isGptModel(modelID) && !isGptNativeSisyphusModel(modelID)) {
showToast(ctx, input.sessionID)
input.agent = resolveRegisteredAgentName("hephaestus") ?? "hephaestus"
if (output?.message) {
+82 -18
View File
@@ -1,4 +1,7 @@
/// <reference types="bun-types" />
import { describe, expect, spyOn, test } from "bun:test"
import type { PluginInput } from "@opencode-ai/plugin"
import { _resetForTesting, updateSessionAgent } from "../../features/claude-code-session-state"
import { getAgentDisplayName } from "../../shared/agent-display-names"
import { createNoSisyphusGptHook } from "./index"
@@ -6,20 +9,29 @@ import { createNoSisyphusGptHook } from "./index"
const SISYPHUS_DISPLAY = getAgentDisplayName("sisyphus")
const HEPHAESTUS_DISPLAY = getAgentDisplayName("hephaestus")
function createOutput() {
type HookOutput = {
message: { agent?: string; variant?: string; [key: string]: unknown }
parts: unknown[]
}
function createOutput(): HookOutput {
return {
message: {},
parts: [],
}
}
function createHookContext(showToast: (input: unknown) => Promise<unknown>): PluginInput {
return {
client: { tui: { showToast } },
} as unknown as PluginInput
}
describe("no-sisyphus-gpt hook", () => {
test("shows toast on every chat.message when sisyphus uses gpt model", async () => {
// given - sisyphus (display name) with gpt model
const showToast = spyOn({ fn: async () => ({}) }, "fn")
const hook = createNoSisyphusGptHook({
client: { tui: { showToast } },
} as any)
const hook = createNoSisyphusGptHook(createHookContext(showToast))
const output1 = createOutput()
const output2 = createOutput()
@@ -40,10 +52,11 @@ describe("no-sisyphus-gpt hook", () => {
expect(showToast).toHaveBeenCalledTimes(2)
expect(output1.message.agent).toBe("hephaestus")
expect(output2.message.agent).toBe("hephaestus")
expect(showToast.mock.calls[0]?.[0]).toMatchObject({
const firstToastCall = (showToast.mock.calls as Array<Array<unknown>>)[0]?.[0]
expect(firstToastCall).toMatchObject({
body: {
title: "NEVER Use Sisyphus with GPT",
message: expect.stringContaining("For GPT models (other than 5.4), always use Hephaestus."),
message: expect.stringContaining("For other GPT models, always use Hephaestus."),
variant: "error",
},
})
@@ -52,9 +65,7 @@ describe("no-sisyphus-gpt hook", () => {
test("does not show toast for gpt-5.4 model (Sisyphus has specialized support)", async () => {
// given - sisyphus with gpt-5.4 model (should be allowed)
const showToast = spyOn({ fn: async () => ({}) }, "fn")
const hook = createNoSisyphusGptHook({
client: { tui: { showToast } },
} as any)
const hook = createNoSisyphusGptHook(createHookContext(showToast))
const output = createOutput()
@@ -70,12 +81,69 @@ describe("no-sisyphus-gpt hook", () => {
expect(output.message.agent).toBeUndefined()
})
test("does not show toast for gpt-5.5 model (native Sisyphus support)", async () => {
// given - sisyphus with gpt-5.5 model (should be allowed)
const showToast = spyOn({ fn: async () => ({}) }, "fn")
const hook = createNoSisyphusGptHook(createHookContext(showToast))
const output = createOutput()
// when - chat.message runs with gpt-5.5
await hook["chat.message"]?.({
sessionID: "ses_gpt55",
agent: SISYPHUS_DISPLAY,
model: { providerID: "openai", modelID: "gpt-5.5" },
}, output)
// then - no toast, agent NOT switched to Hephaestus
expect(showToast).toHaveBeenCalledTimes(0)
expect(output.message.agent).toBeUndefined()
})
test("sets medium variant for gpt-5.5 model when native Sisyphus support is used", async () => {
// given - sisyphus with gpt-5.5 model and no selected variant
const showToast = spyOn({ fn: async () => ({}) }, "fn")
const hook = createNoSisyphusGptHook(createHookContext(showToast))
const output = createOutput()
// when - chat.message runs with gpt-5.5
await hook["chat.message"]?.({
sessionID: "ses_gpt55_medium",
agent: SISYPHUS_DISPLAY,
model: { providerID: "openai", modelID: "gpt-5.5" },
}, output)
// then - Sisyphus stays active and receives its configured GPT-5.5 variant
expect(showToast).toHaveBeenCalledTimes(0)
expect(output.message.agent).toBeUndefined()
expect(output.message.variant).toBe("medium")
})
test("preserves selected variant for gpt-5.5 model when native Sisyphus support is used", async () => {
// given - sisyphus with gpt-5.5 model and a selected variant
const showToast = spyOn({ fn: async () => ({}) }, "fn")
const hook = createNoSisyphusGptHook(createHookContext(showToast))
const output: HookOutput = { message: { variant: "high" }, parts: [] }
// when - chat.message runs with gpt-5.5
await hook["chat.message"]?.({
sessionID: "ses_gpt55_high",
agent: SISYPHUS_DISPLAY,
model: { providerID: "openai", modelID: "gpt-5.5" },
}, output)
// then - user-selected variant is not overwritten
expect(showToast).toHaveBeenCalledTimes(0)
expect(output.message.agent).toBeUndefined()
expect(output.message.variant).toBe("high")
})
test("does not show toast for non-gpt model", async () => {
// given - sisyphus with claude model
const showToast = spyOn({ fn: async () => ({}) }, "fn")
const hook = createNoSisyphusGptHook({
client: { tui: { showToast } },
} as any)
const hook = createNoSisyphusGptHook(createHookContext(showToast))
const output = createOutput()
@@ -94,9 +162,7 @@ describe("no-sisyphus-gpt hook", () => {
test("does not show toast for non-sisyphus agent", async () => {
// given - hephaestus with gpt model
const showToast = spyOn({ fn: async () => ({}) }, "fn")
const hook = createNoSisyphusGptHook({
client: { tui: { showToast } },
} as any)
const hook = createNoSisyphusGptHook(createHookContext(showToast))
const output = createOutput()
@@ -117,9 +183,7 @@ describe("no-sisyphus-gpt hook", () => {
_resetForTesting()
updateSessionAgent("ses_4", SISYPHUS_DISPLAY)
const showToast = spyOn({ fn: async () => ({}) }, "fn")
const hook = createNoSisyphusGptHook({
client: { tui: { showToast } },
} as any)
const hook = createNoSisyphusGptHook(createHookContext(showToast))
const output = createOutput()
@@ -2,6 +2,65 @@ import { describe, expect, test } from "bun:test"
import { injectContinuationPrompt } from "./continuation-prompt-injector"
describe("ralph-loop continuation prompt injector", () => {
test("#given inherited message agent has ZWSP prefix #when injecting continuation prompt #then promptAsync receives normalized agent", async () => {
// given
let promptBody: { agent?: string } | undefined
const ctx = {
client: {
session: {
messages: async () => ({
data: [{ info: { agent: "\u200bSisyphus - Ultraworker" } }],
}),
promptAsync: async (input: { body: { agent?: string } }) => {
promptBody = input.body
return {}
},
},
},
}
// when
await injectContinuationPrompt(ctx as never, {
sessionID: "ses_ralph_zwsp_agent",
prompt: "continue",
directory: "/tmp/test",
apiTimeoutMs: 50,
})
// then
expect(promptBody?.agent).toBe("sisyphus")
expect(promptBody?.agent).not.toContain("\u200b")
})
test("#given inherited message agent has no ZWSP prefix #when injecting continuation prompt #then promptAsync receives normalized agent", async () => {
// given
let promptBody: { agent?: string } | undefined
const ctx = {
client: {
session: {
messages: async () => ({
data: [{ info: { agent: "Sisyphus - Ultraworker" } }],
}),
promptAsync: async (input: { body: { agent?: string } }) => {
promptBody = input.body
return {}
},
},
},
}
// when
await injectContinuationPrompt(ctx as never, {
sessionID: "ses_ralph_clean_agent",
prompt: "continue",
directory: "/tmp/test",
apiTimeoutMs: 50,
})
// then
expect(promptBody?.agent).toBe("sisyphus")
})
test("#given inherited message model includes variant #when injecting continuation prompt #then promptAsync receives variant as a top-level field", async () => {
// given
let promptBody:
@@ -8,6 +8,7 @@ import {
normalizeSDKResponse,
resolveInheritedPromptTools,
} from "../../shared"
import { normalizeAgentForPromptKey } from "../../shared/agent-display-names"
type MessageInfo = {
agent?: string
@@ -69,6 +70,7 @@ export async function injectContinuationPrompt(
}
const inheritedTools = resolveInheritedPromptTools(sourceSessionID, tools)
const cleanAgent = normalizeAgentForPromptKey(agent)
const launchModel = model
? { providerID: model.providerID, modelID: model.modelID }
@@ -78,7 +80,7 @@ export async function injectContinuationPrompt(
await ctx.client.session.promptAsync({
path: { id: options.sessionID },
body: {
...(agent !== undefined ? { agent } : {}),
...(cleanAgent !== undefined ? { agent: cleanAgent } : {}),
...(launchModel ? { model: launchModel } : {}),
...(launchVariant ? { variant: launchVariant } : {}),
...(inheritedTools ? { tools: inheritedTools } : {}),
+42 -20
View File
@@ -17,7 +17,7 @@ describe("ralph-loop", () => {
let mockSessionMessages: Array<{ info?: { role?: string }; parts?: Array<{ type: string; text?: string }> }>
let mockMessagesApiResponseShape: "data" | "array"
function createMockPluginInput() {
function createMockPluginInput(): Parameters<typeof createRalphLoopHook>[0] {
return {
client: {
session: {
@@ -63,7 +63,7 @@ describe("ralph-loop", () => {
},
},
directory: TEST_DIR,
} as unknown as Parameters<typeof createRalphLoopHook>[0]
} as Parameters<typeof createRalphLoopHook>[0]
}
beforeEach(() => {
@@ -304,6 +304,33 @@ describe("ralph-loop", () => {
expect(state?.iteration).toBe(2)
})
test("should skip continuation when background task is running", async () => {
// given - active loop state with a running background task
const hook = createRalphLoopHook(createMockPluginInput(), {
backgroundManager: {
getTasksByParentSession: (sessionID: string) => sessionID === "session-123"
? [{ status: "running" }]
: [],
},
})
hook.startLoop("session-123", "Build a feature", { maxIterations: 10 })
// when - session goes idle
await hook.event({
event: {
type: "session.idle",
properties: { sessionID: "session-123" },
},
})
// then - no continuation should be injected
expect(promptCalls.length).toBe(0)
// then - iteration should not be incremented
const state = hook.getState()
expect(state?.iteration).toBe(1)
})
test("should stop loop when max iterations reached", async () => {
// given - loop at max iteration
const hook = createRalphLoopHook(createMockPluginInput())
@@ -359,8 +386,8 @@ describe("ralph-loop", () => {
expect(hook.getState()).not.toBeNull()
})
test("should skip injection during recovery", async () => {
// given - active loop and session in recovery
test("should continue after non-abort session error", async () => {
// given - active loop and non-abort session error
const hook = createRalphLoopHook(createMockPluginInput())
hook.startLoop("session-123", "Test task")
@@ -371,7 +398,7 @@ describe("ralph-loop", () => {
},
})
// when - session goes idle immediately
// when - session goes idle immediately after the error
await hook.event({
event: {
type: "session.idle",
@@ -379,8 +406,9 @@ describe("ralph-loop", () => {
},
})
// then - no continuation injected
expect(promptCalls.length).toBe(0)
// then - continuation is injected without a recovery skip
expect(promptCalls.length).toBe(1)
expect(hook.getState()?.iteration).toBe(2)
})
test("should clear state on session deletion", async () => {
@@ -1144,20 +1172,14 @@ Original task: Build something`
test("should not hang when session.messages() throws", async () => {
// given - API that throws (simulates timeout error)
let apiCallCount = 0
const errorMock = {
...createMockPluginInput(),
client: {
...createMockPluginInput().client,
session: {
...createMockPluginInput().client.session,
messages: async () => {
apiCallCount++
throw new Error("API timeout")
},
},
const errorMock = createMockPluginInput()
Object.defineProperty(errorMock.client.session, "messages", {
value: async () => {
apiCallCount++
throw new Error("API timeout")
},
}
const hook = createRalphLoopHook(errorMock as any, {
})
const hook = createRalphLoopHook(errorMock, {
getTranscriptPath: () => join(TEST_DIR, "nonexistent.jsonl"),
apiTimeout: 100,
})
@@ -1,33 +0,0 @@
type SessionState = {
isRecovering?: boolean
}
export function createLoopSessionRecovery(options?: { recoveryWindowMs?: number }) {
const recoveryWindowMs = options?.recoveryWindowMs ?? 5000
const sessions = new Map<string, SessionState>()
function getSessionState(sessionID: string): SessionState {
let state = sessions.get(sessionID)
if (!state) {
state = {}
sessions.set(sessionID, state)
}
return state
}
return {
isRecovering(sessionID: string): boolean {
return getSessionState(sessionID).isRecovering === true
},
markRecovering(sessionID: string): void {
const state = getSessionState(sessionID)
state.isRecovering = true
setTimeout(() => {
state.isRecovering = false
}, recoveryWindowMs)
},
clear(sessionID: string): void {
sessions.delete(sessionID)
},
}
}
@@ -0,0 +1,96 @@
/// <reference types="bun-types" />
import { afterEach, beforeEach, describe, expect, test } from "bun:test"
import { existsSync, mkdirSync, rmSync } from "node:fs"
import { tmpdir } from "node:os"
import { join } from "node:path"
import { createRalphLoopHook } from "./index"
import { clearState } from "./storage"
describe("ralph-loop non-abort error continuation", () => {
const testDirectory = join(tmpdir(), `ralph-loop-non-abort-error-${Date.now()}`)
let promptCalls: Array<{ sessionID: string; text: string }>
let messagesCalls: Array<{ sessionID: string }>
beforeEach(() => {
promptCalls = []
messagesCalls = []
mkdirSync(testDirectory, { recursive: true })
clearState(testDirectory)
})
afterEach(() => {
clearState(testDirectory)
if (existsSync(testDirectory)) {
rmSync(testDirectory, { recursive: true, force: true })
}
})
test("continues on next idle after non-abort session error", async () => {
// given - an active Ralph Loop receives a recoverable command error
const hook = createRalphLoopHook({
directory: testDirectory,
project: testDirectory,
worktree: testDirectory,
serverUrl: "http://localhost:4096",
$: async () => ({}),
client: {
session: {
messages: async (options: { path: { id: string } }) => {
messagesCalls.push({ sessionID: options.path.id })
return { data: [] }
},
promptAsync: async (options: {
path: { id: string }
body: { parts: Array<{ type: string; text: string }> }
}) => {
promptCalls.push({
sessionID: options.path.id,
text: options.body.parts[0]?.text ?? "",
})
return {}
},
prompt: async (options: {
path: { id: string }
body: { parts: Array<{ type: string; text: string }> }
}) => {
promptCalls.push({
sessionID: options.path.id,
text: options.body.parts[0]?.text ?? "",
})
return {}
},
},
tui: {
showToast: async () => ({}),
},
},
} as never)
hook.startLoop("session-123", "Keep working", {
messageCountAtStart: 0,
maxIterations: 5,
})
await hook.event({
event: {
type: "session.error",
properties: {
sessionID: "session-123",
error: { name: "CommandFailedError" },
},
},
})
// when - OpenCode emits the idle event caused by that failed command
await hook.event({
event: { type: "session.idle", properties: { sessionID: "session-123" } },
})
// then - the loop should continue instead of skipping idle as recovery
expect(promptCalls).toHaveLength(1)
expect(promptCalls[0]?.sessionID).toBe("session-123")
expect(promptCalls[0]?.text).toContain("Keep working")
expect(messagesCalls.length).toBeGreaterThan(0)
expect(hook.getState()?.iteration).toBe(2)
})
})
@@ -11,11 +11,6 @@ import { continueIteration } from "./iteration-continuation"
import { handlePendingVerification } from "./pending-verification-handler"
import { handleDeletedLoopSession, handleErroredLoopSession } from "./session-event-handler"
type SessionRecovery = {
isRecovering: (sessionID: string) => boolean
markRecovering: (sessionID: string) => void
clear: (sessionID: string) => void
}
type LoopStateController = {
getState: () => RalphLoopState | null
clear: () => boolean
@@ -25,7 +20,7 @@ type LoopStateController = {
setVerificationSessionID: (sessionID: string, verificationSessionID: string) => RalphLoopState | null
restartAfterFailedVerification: (sessionID: string, messageCountAtStart?: number) => RalphLoopState | null
}
type RalphLoopEventHandlerOptions = { directory: string; apiTimeoutMs: number; getTranscriptPath: (sessionID: string) => string | undefined; checkSessionExists?: RalphLoopOptions["checkSessionExists"]; sessionRecovery: SessionRecovery; loopState: LoopStateController }
type RalphLoopEventHandlerOptions = { directory: string; apiTimeoutMs: number; getTranscriptPath: (sessionID: string) => string | undefined; checkSessionExists?: RalphLoopOptions["checkSessionExists"]; backgroundManager?: RalphLoopOptions["backgroundManager"]; loopState: LoopStateController }
export function createRalphLoopEventHandler(
ctx: PluginInput,
@@ -48,14 +43,17 @@ export function createRalphLoopEventHandler(
inFlightSessions.add(sessionID)
try {
if (options.sessionRecovery.isRecovering(sessionID)) {
log(`[${HOOK_NAME}] Skipped: in recovery`, { sessionID })
const state = options.loopState.getState()
if (!state || !state.active) {
return
}
const state = options.loopState.getState()
if (!state || !state.active) {
const hasRunningBackgroundTasks = options.backgroundManager
? options.backgroundManager.getTasksByParentSession(sessionID).some((task: { status: string }) => task.status === "running")
: false
if (hasRunningBackgroundTasks) {
log(`[${HOOK_NAME}] Skipped: background tasks running`, { sessionID })
return
}
@@ -220,12 +218,12 @@ export function createRalphLoopEventHandler(
}
if (event.type === "session.deleted") {
if (!handleDeletedLoopSession(props, options.loopState, options.sessionRecovery)) return
if (!handleDeletedLoopSession(props, options.loopState)) return
return
}
if (event.type === "session.error") {
handleErroredLoopSession(props, options.loopState, options.sessionRecovery)
handleErroredLoopSession(props, options.loopState)
}
}
}
+2 -3
View File
@@ -1,7 +1,6 @@
import type { PluginInput } from "@opencode-ai/plugin"
import type { RalphLoopOptions, RalphLoopState } from "./types"
import { getTranscriptPath as getDefaultTranscriptPath } from "../claude-code-hooks/transcript"
import { createLoopSessionRecovery } from "./loop-session-recovery"
import { createLoopStateController } from "./loop-state-controller"
import { createRalphLoopEventHandler } from "./ralph-loop-event-handler"
@@ -46,20 +45,20 @@ export function createRalphLoopHook(
const getTranscriptPath = options?.getTranscriptPath ?? getDefaultTranscriptPath
const apiTimeout = options?.apiTimeout ?? DEFAULT_API_TIMEOUT
const checkSessionExists = options?.checkSessionExists
const backgroundManager = options?.backgroundManager
const loopState = createLoopStateController({
directory: ctx.directory,
stateDir,
config,
})
const sessionRecovery = createLoopSessionRecovery()
const event = createRalphLoopEventHandler(ctx, {
directory: ctx.directory,
apiTimeoutMs: apiTimeout,
getTranscriptPath,
checkSessionExists,
sessionRecovery,
backgroundManager,
loopState,
})
+1 -10
View File
@@ -7,15 +7,9 @@ type LoopStateController = {
clear: () => boolean
}
type SessionRecovery = {
clear: (sessionID: string) => void
markRecovering: (sessionID: string) => void
}
export function handleDeletedLoopSession(
props: Record<string, unknown> | undefined,
loopState: LoopStateController,
sessionRecovery: SessionRecovery,
): boolean {
const sessionInfo = props?.info as { id?: string } | undefined
if (!sessionInfo?.id) return false
@@ -25,14 +19,12 @@ export function handleDeletedLoopSession(
loopState.clear()
log(`[${HOOK_NAME}] Session deleted, loop cleared`, { sessionID: sessionInfo.id })
}
sessionRecovery.clear(sessionInfo.id)
return true
}
export function handleErroredLoopSession(
props: Record<string, unknown> | undefined,
loopState: LoopStateController,
sessionRecovery: SessionRecovery,
): boolean {
const sessionID = props?.sessionID as string | undefined
const error = props?.error as { name?: string } | undefined
@@ -44,13 +36,12 @@ export function handleErroredLoopSession(
loopState.clear()
log(`[${HOOK_NAME}] User aborted, loop cleared`, { sessionID })
}
sessionRecovery.clear(sessionID)
}
return true
}
if (sessionID) {
sessionRecovery.markRecovering(sessionID)
log(`[${HOOK_NAME}] Session error ignored, loop remains active`, { sessionID })
}
return true
}
+1
View File
@@ -22,4 +22,5 @@ export interface RalphLoopOptions {
getTranscriptPath?: (sessionId: string) => string
apiTimeout?: number
checkSessionExists?: (sessionId: string) => Promise<boolean>
backgroundManager?: { getTasksByParentSession: (sessionId: string) => Array<{ status: string }> }
}
@@ -1053,7 +1053,6 @@ describe("todo-continuation-enforcer", () => {
})
test("should show countdown toast updates", async () => {
fakeTimers.restore()
// given - session with incomplete todos
const sessionID = "main-toast"
setMainSession(sessionID)
@@ -1066,7 +1065,7 @@ describe("todo-continuation-enforcer", () => {
})
// then - multiple toast updates during countdown (2s countdown = 2 toasts: "2s" and "1s")
await wait(2500)
await fakeTimers.advanceBy(1500)
expect(toastCalls.length).toBeGreaterThanOrEqual(2)
expect(toastCalls[0].message).toContain("2s")
}, { timeout: 15000 })