Skip to content

Feat/multi agent orchestration phase1#11

Merged
webdevcom01-cell merged 3 commits into
mainfrom
feat/multi-agent-orchestration-phase1
Mar 28, 2026
Merged

Feat/multi agent orchestration phase1#11
webdevcom01-cell merged 3 commits into
mainfrom
feat/multi-agent-orchestration-phase1

Conversation

@webdevcom01-cell
Copy link
Copy Markdown
Owner

Description

Briefly describe the changes in this PR and the motivation behind them.

Type of change

  • New feature
  • Bug fix
  • Documentation update
  • Refactor (no behavior change)
  • Test update
  • Chore (build, CI, dependencies)

Checklist

  • I have read the Contributing Guide
  • My code follows the project code standards (TypeScript strict, no any, Tailwind only)
  • I have added tests that cover my changes
  • All tests pass (pnpm test)
  • Linting passes (pnpm lint)
  • Type checking passes (pnpm typecheck)
  • I have updated documentation if needed
  • No console.log statements left in code
  • No hardcoded secrets or internal URLs

webdevcom01-cell and others added 3 commits March 28, 2026 19:59
Prevent Dependabot from auto-creating PRs for major version upgrades
(e.g. Prisma 6→7, eslint 9→10, vitest 3→4). Major bumps require manual
review and code updates before merging.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Extend major-version ignore rule to GitHub Actions and Docker in addition
to npm. This prevents PRs for Node Alpine upgrades (20→25) and major
GitHub Actions version bumps since we use SHA-pinned actions.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
…e Loop

Add heterogeneous model routing with plan-and-execute pattern (powerful model
plans, cheap models execute) and self-correcting reflexive loop node. Includes:

- plan_and_execute node: task decomposition with tier-based model selection
- reflexive_loop node: generate → evaluate → retry with feedback loop
- getModelByTier() / getModelFallbackChain() for dynamic model routing
- Adaptive cost monitor mode (auto-downgrades tier at 60/80/95% budget)
- LLM-based meta-orchestrator upgrade (keyword fallback preserved)
- Flow builder UI: node components, property editors, node picker entries
- Starter flow template: Plan-and-Execute Pipeline
- Flow validator updated with all 55 node types
- Implementation plan document in docs/

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
@vercel
Copy link
Copy Markdown

vercel Bot commented Mar 28, 2026

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

Project Deployment Actions Updated (UTC)
agent-studio Ready Ready Preview, Comment Mar 28, 2026 9:48pm

Request Review

@webdevcom01-cell webdevcom01-cell merged commit 8e14377 into main Mar 28, 2026
9 of 10 checks passed
@railway-app railway-app Bot temporarily deployed to reliable-youth / production March 28, 2026 22:00 Inactive
@webdevcom01-cell webdevcom01-cell deleted the feat/multi-agent-orchestration-phase1 branch March 29, 2026 22:23
webdevcom01-cell added a commit that referenced this pull request May 24, 2026
- Migration 20260605000000: enable RLS + FORCE on PolicyDecision,
  4 policies for app_user, organizationId+id composite index
- approval-engine: wrap requestApproval findFirst+create, resolveDecision
  findUnique+update, waitForDecision findUnique in withOrgContext;
  add organizationId param to resolveDecision and waitForDecision;
  processTimeouts stays unwrapped (cross-org cron, BYPASSRLS)
- decisions/[decisionId]/route: loadDecision to withAdminBypass,
  GET findUnique and DELETE update to withOrgContext, pass organizationId
  to resolveDecision
- policies/[policyId]/route: policyDecision.count to withOrgContext
- policies/[policyId]/decisions/route: findMany to withOrgContext
- pending-approvals/route: agent lookup for orgId, findMany to withOrgContext
- approval-engine tests: update resolveDecision and waitForDecision call
  signatures to match new organizationId param positions
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant