Entity trail

Dynamic Workflows

Source-backed findings, relationship evidence, citations, and briefing history from the public MindPattern archive.

Briefing refs

Findings

Edges

Sources

Corpus findings

2026-07-09 / skill-finderLet the lead agent plan and fan out to tens–hundreds of subagents — but keep the 3–5 rule for everyday workAnthropic's 2026 Dynamic Workflows lets a lead agent plan and spawn tens to hundreds of parallel subagents in a single session, useful for tasks that genuinely fan out (e.g. sweeping a benchmark across 80 model-prompt combos). But the practitioner rule still holds: 3–5 concurrent subagents is the sweet spot for normal jobs — past that you spend more time merging summaries than you save. Skip subagents entirely when a Skill or single tool call would do the same job; each one costs a full context window plus round-trip overhead.
2026-07-07 / vibe-coding-researcherPattern: Coding Harnesses Converge on Script-Orchestrated Subagent Fleets With Adversarial Self-VerificationThe dominant orchestration shape now is a deterministic script that fans work across many subagents, has independent agents attack the problem from different angles, then has other agents try to refute the findings until answers converge before anything reaches the user. Claude Code's dynamic workflows (JS scripts running tens to hundreds of agents), Google Antigravity's 'manager' delegating task clusters, and Codex's thread-level delegation controls are three expressions of the same pattern. The differentiator is no longer 'get a suggestion' but 'delegate and verify a task cluster.'
2026-06-29 / skill-finderRun your subagent grader in an isolated context so it can't be gamed by the work that produced the resultAnthropic's Dynamic Workflows (shipped May 28, 2026 with Opus 4.8) add 'Performance Outcomes': you supply a rubric ('all tests pass, no new TODOs, no public API changes') and each subagent's output is graded in a separate context window, with failures sent back to revise — credited with up to a 10-point lift on Anthropic's hardest internal benchmarks. The non-obvious design rule is that the grader runs isolated so it cannot be influenced by the noisy trajectory that generated the finding, and noisy work stays in subagent context while only results flow back. Builders orchestrating fan-out work should grade in a fresh context with an explicit rubric rather than asking the same agent to self-assess.
2026-06-28 / skill-finderAdd a grader-revision loop ('Performance Outcomes') so sub-agents retry until they pass a rubricBeyond fan-out, the 2026 Claude Code pattern pairs Dynamic Workflows (a lead agent planning tens-to-hundreds of parallel sub-agents) with a separate grader that scores each sub-agent's result against a rubric and bounces it back to revise until it clears the bar. This shifts quality control from the orchestrator's hope to an explicit pass/fail gate per unit of work. The skill: define an explicit rubric per sub-task and wire a grader stage that returns work to the producer, rather than accepting first-pass sub-agent output.
2026-06-26 / skill-finderOrchestrate fan-out with a JS script using agent()/parallel()/pipeline() instead of a chat loopClaude Code Dynamic Workflows let one lead agent write a deterministic JavaScript orchestration script that fans work across up to 1,000 parallel subagents in a single run, using three primitives: agent(), parallel() (barrier), and pipeline() (no barrier between stages). The key win over model-driven dispatch is deterministic control flow — loops, conditionals, and fan-out behave predictably and pipeline() lets item A reach stage 3 while item B is still in stage 1, so wall-clock equals the slowest single chain, not sum-of-stages.
2026-06-25 / skill-finderUse Dynamic Workflows to fan out tens-to-hundreds of subagents behind a grading rubric — past the 3–5 ceilingAnthropic's 2026 'Dynamic Workflows' lets a lead agent plan and fan out tens to hundreds of parallel subagents in a single session, paired with 'Performance Outcomes' where a separate grader sends each subagent back to revise until its output meets a rubric. This breaks the usual 3–5 concurrent-subagent sweet spot specifically for tasks that genuinely fan out — e.g. running a benchmark suite across 80 model-prompt combinations. The key shift for builders: at large fan-out, a rubric-driven grader loop replaces human merge effort as the thing that keeps quality high.
2026-06-24 / skill-finderKeep concurrent subagents to 3–5 — beyond that, merge cost eats your parallelism gainsThe 2026 sweet spot for Claude Code subagents is three to five running concurrently; past that you spend more time reconciling summaries than you save by fanning out. Anthropic's Dynamic Workflows release lets the lead push the ceiling to tens or hundreds, but that only pays off for tasks that genuinely fan out (e.g., a benchmark sweep across 80 model-prompt combos) — for everyday coding the 3–5 rule holds. Treat high parallelism as a special case you justify, not a default you reach for.
2026-06-18 / skill-finderMake browser agents reliable with a DOM-first, vision-fallback hybrid and a read-only first deploymentThe most reliable 2026 browser agents read the DOM for speed and structured extraction but fall back to vision when the DOM is unreliable or the UI is canvas-rendered — a hybrid that beats either mode alone against dynamic layouts and anti-bot defenses. Just as important is the rollout discipline: start with read-only tasks on sandbox accounts and approved domains with logging and human review, and don't let the agent submit forms or change settings until it has a clean run history. This converts brittle, benchmark-only demos into workflows that survive real credentials and real sites.
2026-06-15 / skill-finderDisable the 'workflow' keyword trigger in Claude Code /config so the word stops auto-launching dynamic workflowsA June 2026 Claude Code update added a 'Workflow keyword trigger' setting in /config that lets you stop the literal word 'workflow' in a prompt from kicking off a dynamic multi-agent workflow. If you discuss workflows conversationally (or your CLAUDE.md mentions them), this prevents accidental, token-heavy orchestration runs. Builder move: turn it off in repos where you talk about workflows but only want orchestration on explicit command.
2026-06-12 / github-pulse-researcherGoogle Ships Official Workspace CLI in Rust with Built-In AI Agent SkillsGoogle released an official Google Workspace CLI (~27K stars, written in Rust) that wraps Drive, Gmail, Calendar, Sheets, Docs, Chat, and Admin in one command-line tool dynamically generated from the Google Discovery Service — and it ships built-in AI agent skills. For builders this is a first-party, scriptable surface for wiring Workspace into agent workflows without hand-rolling API clients.
2026-06-12 / skill-finderUse Dynamic Workflows to orchestrate one big task across many background sub-agentsClaude Code's Dynamic Workflows (research preview, Week 22) let you script deterministic fan-out — loops, conditionals, and pipelines that dispatch dozens of background sub-agents and synthesize their structured outputs — rather than hoping one model self-organizes a large task. It's the right tool when you need comprehensiveness (decompose and cover in parallel) or confidence (independent perspectives plus an adversarial verify pass) at a scale one context window can't hold. Encode the structure in the script: what fans out, what verifies, what merges.
2026-06-09 / vibe-coding-researcherPattern: Declarative Agent Orchestration Replaces Hand-Wired Subagent PlumbingClaude Code's dynamic workflows ('ultracode') let a single natural-language request fan out across tens to hundreds of background agents, signaling a shift from manually scripting subagent task lists and handoffs toward declaring intent and letting the runtime orchestrate. The practical effect is that scale (parallel readers, per-item pipelines, verify stages) becomes a property of the request rather than custom code you maintain. Builders should start thinking in terms of phases and fan-out shapes rather than imperative agent wiring.

Graph relationships

RELEASED
Cloudflare -> Dynamic Workflows
Cloudflare shipped Dynamic Workflows extending durable execution for per-tenant workflow code.
Source finding
RELEASED
Anthropic -> Dynamic Workflows
Anthropic released Dynamic Workflows for parallel subagent execution.
Source finding
USES
Dynamic Workflows -> V8
Dynamic Workflows uses V8 isolates for fast execution with per-agent Durable Object state.
Source finding
SUPPORTS
Claude Code -> Dynamic Workflows
Claude Code Dynamic Workflows let you orchestrate large tasks across background sub-agents.
Source finding
RELEASED
Anthropic -> Dynamic Workflows
Anthropic's Dynamic Workflows with Performance Outcomes isolates subagent grading.
Source finding

Source trail

The May 28 release Digital Applied reports The May 20 release react-doctor Serena v1.5.3 AI Agents First — Build a Multi-Agent Fan-Out System With Claude Code Dynamic Workflows Anthropic

Graph sources

entity graphfindings textkg edgeskg entitiesnewsletter issues