refactor delegation system, add PR summary comments, and improve code quality (#334)
* refactor delegation system and add PR summary comments Delegation system: - replace mode-based delegation with select_mode → delegate two-step flow - orchestrator crafts self-contained subagent prompts (clean context — no system/repo/event instructions leak) - add role-based tool filtering via FastMCP authenticate hook (?role=subagent hides orchestrator-only tools) - add select_mode tool for orchestrator guidance per mode - add ask_question tool for lightweight research subagents - extract shared subagent lifecycle into subagent.ts (create, complete, stdout, instructions) - route set_output to per-subagent state when activeSubagentId is set - track per-subagent state (SubagentState Map) replacing boolean delegationActive flag - capture and aggregate AgentUsage across all agents (claude, codex, gemini, opencode) - write usage summary table to GitHub job summary - block built-in subagent spawning (Task for Claude, Task(*) for Cursor) - increase activity timeout from 60s to 300s (subagent thinking phases) - fix gh CLI misguidance in system prompt — explicitly forbid usage PR summary comments: - add prSummaryComment trigger (DB schema + migrations + Zod + UI toggle) - dispatch mini-effort summary job alongside PR review on pr.created - add update_pull_request_body MCP tool - add defaultEffort option to webhook dispatch Hardening: - rewrite delegate/selectMode tests with simulated state management - add toolFiltering.test.ts for role extraction, canAccess, set_output routing - remove non-null assertions for PULLFROG_TEMP_DIR (proper error throws) - use fetchWithRetry for direct tarball downloads - DRY fix for rate limit check in test runner Co-authored-by: Cursor <cursoragent@cursor.com> * fix: add type keyword to Effort import in handleWebhook.ts Co-authored-by: Cursor <cursoragent@cursor.com> * clean up delegation system, improve code quality across the codebase - simplify delegate tool to instructions + effort params with subagent lifecycle in subagent.ts - add select_mode and ask_question orchestrator-only tools with canAccess filtering - replace delegate.test.ts/selectMode.test.ts with toolFiltering.test.ts (live MCP integration) - add set_output routing for subagent context and AgentUsage tracking across all agents - add PR summary comment trigger (schema, UI, webhook dispatch with silent flag) - add update_pull_request_body MCP tool - fix changed-agents.sh to always include claude canary for non-agent action changes - fix cursor pagination bug in getSelectedInstallationReposPage - remove destructuring patterns, inline type definitions, and unsafe type casts - replace non-null assertions with explicit checks in install.ts - convert multi-param functions to single param objects (postCleanup, runActionLocal, etc.) - use isHttpError helper in API routes instead of catch-any patterns - add adhoc test fixtures for delegation scenarios (context isolation, error handling, synthesis, etc.) Co-authored-by: Cursor <cursoragent@cursor.com> * no subagent mutation, one mcp per subagent * address review feedback: parallel-safe usage tracking, subagent isolation, minor improvements * fix subagent state isolation: replace Object.freeze with shallow copy Object.freeze throws TypeErrors when subagent tools (checkout_pr, report_progress) write scalar properties to toolState. A shallow copy achieves the same isolation for scalar fields while allowing tools to work normally. Shared references (subagents Map, usageEntries array) remain shared for coordination. --------- Co-authored-by: Cursor <cursoragent@cursor.com> Co-authored-by: pullfrog[bot] <226033991+pullfrog[bot]@users.noreply.github.com>
This commit is contained in:
David Blass
committed by
pullfrog[bot]
pullfrog[bot]
parent
a90743e9fe
commit
cfd38d82fc
+31
-14
@@ -12,7 +12,7 @@ import { log } from "../utils/cli.ts";
|
||||
import { installFromNpmTarball } from "../utils/install.ts";
|
||||
import { spawn } from "../utils/subprocess.ts";
|
||||
import { ThinkingTimer } from "../utils/timer.ts";
|
||||
import { type AgentRunContext, agent } from "./shared.ts";
|
||||
import { type AgentRunContext, type AgentUsage, agent } from "./shared.ts";
|
||||
|
||||
// model selection based on effort level
|
||||
// these are aliases that always resolve to the latest version
|
||||
@@ -40,10 +40,11 @@ function buildDisallowedTools(ctx: AgentRunContext): string[] {
|
||||
// both "disabled" and "restricted" block native bash
|
||||
// "restricted" means use MCP bash tool instead
|
||||
const bash = ctx.payload.bash;
|
||||
if (bash !== "enabled") disallowed.push("Bash", "Task(Bash)");
|
||||
if (bash !== "enabled") disallowed.push("Bash");
|
||||
// always block native file tools (use MCP file_read/file_write instead)
|
||||
disallowed.push("Read", "Write", "Edit", "MultiEdit");
|
||||
disallowed.push("Task(Read)", "Task(Write)", "Task(Edit)", "Task(MultiEdit)");
|
||||
// block built-in subagent spawning — delegation is handled by gh_pullfrog/delegate
|
||||
disallowed.push("Task");
|
||||
return disallowed;
|
||||
}
|
||||
|
||||
@@ -128,6 +129,7 @@ export const claude = agent({
|
||||
|
||||
let stdoutBuffer = "";
|
||||
let finalOutput = "";
|
||||
const usageContainer: UsageContainer = { value: null };
|
||||
|
||||
// Track bash tool IDs to identify when bash tool results come back
|
||||
const bashToolIds = new Set<string>();
|
||||
@@ -139,7 +141,7 @@ export const claude = agent({
|
||||
cwd: process.cwd(),
|
||||
env: process.env,
|
||||
stdio: ["ignore", "pipe", "pipe"],
|
||||
activityTimeout: 0, // disabled: process-level timeout in main.ts handles this (subprocess timeout would kill orchestrator during delegation)
|
||||
activityTimeout: 0, // process-level activity timeout (5min) is the single authority
|
||||
onStdout: async (chunk) => {
|
||||
finalOutput += chunk;
|
||||
markActivity(); // reset activity timeout on any CLI output
|
||||
@@ -162,7 +164,7 @@ export const claude = agent({
|
||||
|
||||
const handler = messageHandlers[message.type];
|
||||
if (handler) {
|
||||
await handler(message as never, bashToolIds, thinkingTimer);
|
||||
await handler(message as never, bashToolIds, thinkingTimer, usageContainer);
|
||||
}
|
||||
} catch {
|
||||
// ignore parse errors - might be non-JSON output
|
||||
@@ -190,6 +192,7 @@ export const claude = agent({
|
||||
success: false,
|
||||
error: errorMessage,
|
||||
output: finalOutput || result.stdout || "",
|
||||
usage: usageContainer.value ?? undefined,
|
||||
};
|
||||
}
|
||||
|
||||
@@ -198,16 +201,21 @@ export const claude = agent({
|
||||
return {
|
||||
success: true,
|
||||
output: finalOutput || result.stdout || "",
|
||||
usage: usageContainer.value ?? undefined,
|
||||
};
|
||||
},
|
||||
});
|
||||
|
||||
// run-local usage container — passed to handlers via closure for parallel-safe runs
|
||||
type UsageContainer = { value: AgentUsage | null };
|
||||
|
||||
type SDKMessageType = SDKMessage["type"];
|
||||
|
||||
type SDKMessageHandler<type extends SDKMessageType = SDKMessageType> = (
|
||||
data: Extract<SDKMessage, { type: type }>,
|
||||
bashToolIds: Set<string>,
|
||||
thinkingTimer: ThinkingTimer
|
||||
thinkingTimer: ThinkingTimer,
|
||||
usageContainer: UsageContainer
|
||||
) => void | Promise<void>;
|
||||
|
||||
type SDKMessageHandlers = {
|
||||
@@ -215,7 +223,7 @@ type SDKMessageHandlers = {
|
||||
};
|
||||
|
||||
const messageHandlers: SDKMessageHandlers = {
|
||||
assistant: (data, bashToolIds, thinkingTimer) => {
|
||||
assistant: (data, bashToolIds, thinkingTimer, _usageContainer) => {
|
||||
if (data.message?.content) {
|
||||
for (const content of data.message.content) {
|
||||
if (content.type === "text" && content.text?.trim()) {
|
||||
@@ -235,7 +243,7 @@ const messageHandlers: SDKMessageHandlers = {
|
||||
}
|
||||
}
|
||||
},
|
||||
user: (data, bashToolIds, thinkingTimer) => {
|
||||
user: (data, bashToolIds, thinkingTimer, _usageContainer) => {
|
||||
if (data.message?.content) {
|
||||
for (const content of data.message.content) {
|
||||
if (typeof content === "string") {
|
||||
@@ -283,7 +291,7 @@ const messageHandlers: SDKMessageHandlers = {
|
||||
}
|
||||
}
|
||||
},
|
||||
result: async (data) => {
|
||||
result: async (data, _bashToolIds, _thinkingTimer, usageContainer) => {
|
||||
if (data.subtype === "success") {
|
||||
const usage = data.usage;
|
||||
const inputTokens = usage?.input_tokens || 0;
|
||||
@@ -292,6 +300,15 @@ const messageHandlers: SDKMessageHandlers = {
|
||||
const outputTokens = usage?.output_tokens || 0;
|
||||
const totalInput = inputTokens + cacheRead + cacheWrite;
|
||||
|
||||
usageContainer.value = {
|
||||
agent: "claude",
|
||||
inputTokens: totalInput,
|
||||
outputTokens,
|
||||
cacheReadTokens: cacheRead,
|
||||
cacheWriteTokens: cacheWrite,
|
||||
costUsd: data.total_cost_usd ?? undefined,
|
||||
};
|
||||
|
||||
log.table([
|
||||
[
|
||||
{ data: "Cost", header: true },
|
||||
@@ -316,9 +333,9 @@ const messageHandlers: SDKMessageHandlers = {
|
||||
log.info(`Failed: ${JSON.stringify(data)}`);
|
||||
}
|
||||
},
|
||||
system: () => {},
|
||||
stream_event: () => {},
|
||||
tool_progress: () => {},
|
||||
tool_use_summary: () => {},
|
||||
auth_status: () => {},
|
||||
system: (_data, _bashToolIds, _thinkingTimer, _usageContainer) => {},
|
||||
stream_event: (_data, _bashToolIds, _thinkingTimer, _usageContainer) => {},
|
||||
tool_progress: (_data, _bashToolIds, _thinkingTimer, _usageContainer) => {},
|
||||
tool_use_summary: (_data, _bashToolIds, _thinkingTimer, _usageContainer) => {},
|
||||
auth_status: (_data, _bashToolIds, _thinkingTimer, _usageContainer) => {},
|
||||
};
|
||||
|
||||
Reference in New Issue
Block a user