✨ refactor(agent-signal): add executeViaExecAgent self-iteration shell by arvinxx · Pull Request #15192 · lobehub/lobehub

arvinxx · 2026-05-25T04:18:47Z

💻 Change Type

✨ feat
✅ test

🔗 Related Issue

Part of LOBE-9434 (parent), LOBE-9454.

Supersedes the Phase 2 piece of #15116 (that PR will be closed; see comment there for context).

🔀 Description of Change

Phase 2 of LOBE-9434: introduces executeViaExecAgent as the async-queue
replacement for executeSelfIteration. Builds on the Phase 1 plumbing
landed in #15187.

Not wired into any caller yet — the review / reflection / feedback
handlers continue to call executeSelfIteration directly. Migration of
each caller happens in Phase 3, one at a time, so each cutover is
independently revertable.

What it does:

Builds the self-iteration prompt via the shared
createAgentSignalSelfIterationPrompt
Routes through AiAgentService.execAgent with the self-iteration
builtin slug and suppressSignal: true so the run cannot recurse
into AgentSignal (shouldSuppressSignal from Phase 1 enforces this)
Caller-supplied window overrides the context-derived defaults

What it does NOT do:

Does not touch executeSelfIteration in execute.ts
Does not change any registered source/signal/action handler
Does not affect any UI or cron path

🧪 How to Test

Tested locally — bun run type-check clean
Added 5 unit tests covering happy path, default mode, window
precedence, context fallback, and enqueue-failure surfacing

bunx vitest run --silent='passed-only' 'executeViaExecAgent.test.ts'

📝 Additional Information

Safe to merge: zero callers reference the new function until Phase 3,
so this is dormant code on the day it lands. Cloud impact: none.

Next: Phase 3 F (userMemory.ts migration to execAgent) — the
high-risk piece — will be a separate dedicated PR so it can be reviewed
and reverted on its own.

Phase 2 of LOBE-9434: introduces `executeViaExecAgent` as the async-queue replacement for `executeSelfIteration` (LOBE-9454). - Builds the self-iteration prompt via the shared `createAgentSignalSelfIterationPrompt` - Routes through `AiAgentService.execAgent` with the `self-iteration` builtin slug and `suppressSignal: true` so the run cannot recurse into AgentSignal - Caller-supplied window overrides the context-derived defaults Not yet wired into any caller — the review / reflection / feedback handlers keep calling `executeSelfIteration` directly. Migration of each caller happens in Phase 3. Tests: 5 unit tests covering happy path, default mode, window precedence, context fallback, and enqueue-failure surfacing.

vercel · 2026-05-25T04:18:53Z

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
lobehub	Ready	Preview, Comment	May 25, 2026 4:49am

sourcery-ai

Sorry @arvinxx, you have reached your weekly rate limit of 500000 diff characters.

Please try again later or upgrade to continue using Sourcery

chatgpt-codex-connector

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: e560c34aba

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

Open a pull request for review
Mark a draft as ready
Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

chatgpt-codex-connector · 2026-05-25T04:23:39Z

+    autoStart: true,
+    maxSteps: input.maxSteps,
+    prompt,
+    slug: BUILTIN_AGENT_SLUGS.selfIteration,


Ensure self-iteration tools are actually available

This enqueues the builtin self-iteration agent by slug, but does not provide the agent-signal-self-iteration tool manifest/executor that the agent depends on. In this repo, that identifier is referenced by the builtin agent runtime (packages/builtin-agents/src/agents/self-iteration/index.ts) but is not registered in the builtin tools list (packages/builtin-tools/src/index.ts only wires selfFeedbackIntentManifest from that package), and AgentToolsEngine drops unknown tool IDs. As a result, once callers migrate to this helper, runs can start without the self-iteration read/write tool surface, so proposal/write behavior will silently fail to execute.

Useful? React with 👍 / 👎.

codecov · 2026-05-25T04:24:28Z

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 70.92%. Comparing base (3c52998) to head (e560c34).

Additional details and impacted files

@@            Coverage Diff            @@
##           canary   #15192     +/-   ##
=========================================
  Coverage   70.92%   70.92%             
=========================================
  Files        3152     3153      +1     
  Lines      314060   314099     +39     
  Branches    28568    34277   +5709     
=========================================
+ Hits       222738   222777     +39     
  Misses      91154    91154             
  Partials      168      168

Flag	Coverage Δ
app	`61.79% <100.00%> (+<0.01%)`	⬆️
database	`92.20% <ø> (ø)`
packages/agent-runtime	`80.48% <ø> (ø)`
packages/builtin-tool-lobe-agent	`19.87% <ø> (ø)`
packages/context-engine	`84.13% <ø> (ø)`
packages/conversation-flow	`91.28% <ø> (ø)`
packages/file-loaders	`87.89% <ø> (ø)`
packages/memory-user-memory	`74.99% <ø> (ø)`
packages/model-bank	`99.99% <ø> (ø)`
packages/model-runtime	`83.88% <ø> (ø)`
packages/prompts	`72.54% <ø> (ø)`
packages/python-interpreter	`92.90% <ø> (ø)`
packages/ssrf-safe-fetch	`0.00% <ø> (ø)`
packages/types	`35.07% <ø> (ø)`
packages/utils	`88.47% <ø> (ø)`
packages/web-crawler	`88.08% <ø> (ø)`

Flags with carried forward coverage won't be shown. Click here to find out more.

Components	Coverage Δ
Store	`68.02% <ø> (ø)`
Services	`54.68% <ø> (ø)`
Server	`72.19% <100.00%> (+0.01%)`	⬆️
Libs	`56.77% <ø> (ø)`
Utils	`85.96% <ø> (ø)`

🚀 New features to boost your workflow:

❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

arvinxx · 2026-05-25T08:59:54Z

Closing this PR — superseded by a different architectural direction.

Context: The shell here was built around a single self-iteration slug, inheriting the consolidation done in commit 627f899 (which itself was reverted into one slug from the originally-three-slug design of PR #15116's early commits). On deeper review, that consolidation conflicted with the standard builtin tool contract:

agent-signal-self-iteration declared one plugin identifier, but its runtime behavior dispatched on mode (review / reflection / feedback) via adapter closures — different reserveOperation / writeReceipt / preflight per mode. Same identifier, three behaviors.
This was a deliberate design choice in PR ♻️ refactor(agent-signal,prompts,database,builtin-tool-self-iteration): unified structure of service, unified tool, unified name and concepts #14699 ("unified structure / unified tool / unified name and concepts"), with no recorded rationale in the PR body. Inferred motivations: code dedup in tools/shared.ts, a single ergonomic surface, easy mode addition.
It worked only because executeSelfIteration bypasses plugin lookup (manually constructing AgentState.operationToolSet). Once we try to route through execAgent (standard plugin lookup), one identifier = one behavior becomes a load-bearing contract that the consolidation violates.
Per-mode receipt tables, idempotency Redis namespaces, brief projection paths are genuinely independent — they're three audit pipelines, not one.

New direction (see follow-up PR):

Restore 3 agent slugs: nightly-review, self-reflection, self-feedback-intent
Each declares its own tool identifier: agent-signal-review, agent-signal-reflection, agent-signal-feedback-intent
Shared scaffolding (tools/shared.ts's idempotency / preflight / receipt span) extracted to a shared module — PR ♻️ refactor(agent-signal,prompts,database,builtin-tool-self-iteration): unified structure of service, unified tool, unified name and concepts #14699's dedup value preserved
Each tool gets its own ExecutionRuntime constructor-injected with mode-specific service (one-shot at boot, matching the selfFeedbackIntent.ts pattern)
Per-execution sourceId + evidenceRefs flow through appContext, the only state still threaded per call

Next PR will start by restoring the 3 slugs (dormant agents with empty plugins). Tool package implementation comes after.

dosubot Bot added the size:M This PR changes 30-99 lines, ignoring generated files. label May 25, 2026

sourcery-ai Bot reviewed May 25, 2026

View reviewed changes

dosubot Bot added the feature:agent Assistant/Agent configuration and behavior label May 25, 2026

arvinxx mentioned this pull request May 25, 2026

feat(agent-signal): converge agent execution onto execAgent; reduce signal layer to event dispatcher [LOBE-9434] #15116

Closed

chatgpt-codex-connector Bot reviewed May 25, 2026

View reviewed changes

arvinxx changed the title ~~✨ feat(agent-signal): add executeViaExecAgent self-iteration shell [LOBE-9434 Phase 2]~~ ✨ feat(agent-signal): add executeViaExecAgent self-iteration shell May 25, 2026

arvinxx changed the title ~~✨ feat(agent-signal): add executeViaExecAgent self-iteration shell~~ ✨ refactor(agent-signal): add executeViaExecAgent self-iteration shell May 25, 2026

vercel Bot deployed to Preview May 25, 2026 04:49 View deployment

arvinxx closed this May 25, 2026

arvinxx mentioned this pull request May 25, 2026

♻️ refactor(agent-signal): restore 3 mode-specific self-iteration agent slugs #15202

Merged

4 tasks

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

✨ refactor(agent-signal): add executeViaExecAgent self-iteration shell#15192

✨ refactor(agent-signal): add executeViaExecAgent self-iteration shell#15192
arvinxx wants to merge 1 commit into
canaryfrom
feat/agent-signal-execute-via-exec-agent

arvinxx commented May 25, 2026

Uh oh!

vercel Bot commented May 25, 2026 •

edited

Loading

Uh oh!

sourcery-ai Bot left a comment

Uh oh!

chatgpt-codex-connector Bot left a comment

Uh oh!

chatgpt-codex-connector Bot May 25, 2026

Uh oh!

codecov Bot commented May 25, 2026 •

edited

Loading

Uh oh!

arvinxx commented May 25, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

Uh oh!

Conversation

arvinxx commented May 25, 2026

💻 Change Type

🔗 Related Issue

🔀 Description of Change

🧪 How to Test

📝 Additional Information

Uh oh!

vercel Bot commented May 25, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

sourcery-ai Bot left a comment

Choose a reason for hiding this comment

Uh oh!

chatgpt-codex-connector Bot left a comment

Choose a reason for hiding this comment

💡 Codex Review

Uh oh!

chatgpt-codex-connector Bot May 25, 2026

Choose a reason for hiding this comment

Uh oh!

codecov Bot commented May 25, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

arvinxx commented May 25, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

vercel Bot commented May 25, 2026 •

edited

Loading

codecov Bot commented May 25, 2026 •

edited

Loading