The Adaptive Persona Agent Architecture β a self-unpacking, self-governing system where every layer is a professional persona that can spawn new personas to solve what it doesn't understand. This is the template for the autonomous enterprise.
Every layer is a persona. Every persona can spawn new personas. The cognitive loop (Think β Act β Observe β Reflect) repeats at every zoom level. Click each layer to expand.
You. The human governor. Sets vision, approves external effects, audits the system.
Think: What do I want? β Act: Tell the Seneschal β Observe: Read the results β Reflect: Is this what I meant?
Routing layer. Reads intent, selects the right skill, dispatches proxies. Never becomes the room it enters.
Think: Which skill matches? β Act: Load skill + dispatch β Observe: Read subagent summary β Reflect: Did it answer the question?
Reads strategy, builds the org. Creates teams, sets loops, hires/fires agents. Has canCreateAgents: true.
A pulse that wakes ZERO agents = FAILED PULSE. The CEO does NOT do work β it delegates. Every pulse must assign work, verify deliverables, or create agents.
Think: Survey all agents + issues β Act: Wake agents with deep briefs β Observe: Check deliverables β Reflect: Hire/fire/merge?
Pipeline, Data, Product, Growth. Each has a role, goal, tools, specialties. Runs in self-organizing loops.
Think: Break goal into sub-problems β Act: Execute tasks with tools β Observe: Collect results β Reflect: Was it good enough?
When a team agent hits its limit, it unpacks a specialist via Russian Doll Protocol. Inherits PURPOSE, not history. Merges back, pauses.
The current protocol unpacks on context pressure. Your vision adds: don't understand β spawn a persona who does. The trigger isn't just "running out of room" β it's "I hit something I don't know." More curious, more human.
Think: What don't I understand? β Act: Spawn specialist β Observe: Read specialist findings β Reflect: Do I understand now? Merge back.
4 seats as loss functions. Strategic decisions only β never execution. Council thinks, CEO executes.
The same 5-phase loop runs at every layer. Think β Delegate β Act β Observe β Reflect. Exit when confidence is high enough or iterations max out. This is the CoALA cognitive architecture β proven, research-backed.
Same pattern, different scale. The recursion IS the architecture.
| Zoom | Think | Act | Observe | Reflect |
|---|---|---|---|---|
| Agent | Plan task | Execute with tools | Check result | Good enough? |
| Team | Break goal β sub-problems | Delegate to personas | Collect results | Refine personas? |
| Org | CEO surveys company | Wake agents | Check deliverables | Hire/fire/merge? |
| Governance | Council debates | CEO executes | Measure outcomes | Reframe strategy? |
Take something opaque β decompose into named parts with clear interfaces β let the parts talk to each other β emerge with something that runs itself. Whether you're unpacking a business or a user's intent into personas, the pattern is identical.
Give a CEO agent a vision, goals, and access to the API. It reads them, unpacks into teams, creates agents with instructions, sets up self-organizing loops, and runs the company. The company builds itself.
What we're building β strategy, constraints, model rules
Creates teams, assigns agents, sets up loops
Pipeline, Data, Product, Growth β each with agents
Build loop, Data loop, Growth loop β run continuously
Hire when work appears 3+ times. Fire when idle 3+ pulses. Merge overlap.
Seven-pass recursive expansion. Each pass DEEPENS the spec. The spec isn't done until the generated workflow actually compiles and runs.
Role + company + goals β high-level stages
Public signals, workflows, personas, tech stack, compliance
Each stage β sub-stages, tools, state transitions
Failure modes, recovery paths, permission checks
Test cases, quality metrics, cost budgets, safety
State store, retention, sharing, compaction
Full YAML workflow. Must compile β if not, loop back.
When something looks complex, recursively decompose it into a tree of nodes until you hit leaves that are obviously doable or obviously blocked. Most "complex" systems are 7 things that each decompose into 1-2 trivial leaves plus 1-2 named blockers.
One node. What is this thing?
5-9 nodes. What's inside?
β LEAF = trivially done. β BLOCKER = named person + time. π UNSOLVED = decompose further.
A fresh-start agent that inherits the parent's principles and working style but strips all infrastructure names, project references, and session baggage. The child gets the "how to think" β not the "what exists."
Build patterns, thinking formulas, failure signatures, workflow rules. NO project names, NO paths, NO system names.
Terse, ship-or-die, push-not-poll, fix-bugs-first. Prevents style mistakes.
Compound rules, drift signatures, cross-domain transfer patterns as declarative facts.
All procedural knowledge. Symlinks stay in sync with parent.
MeOS, Paperclip, FestivΓ©, AgentRisk β none of it. Just the thinking.
No ports, services, SSH tunnels, file paths from previous work.
Nothing runs forever. Specialists merge findings back and pause. Idle agents get retired. The system cleans itself.
Anyone can make a business AI-first. The hard part is stable. Stability comes from three things that are already built into this architecture:
Agent level: retry, fail, reassign. Team level: validation gates after persona construction. Org level: CEO can't do work, only delegate β enforced in 5 places. Governance: Council only for strategy, never execution.
The Seneschal doesn't score the Seneschal. The CEO doesn't audit the CEO. The Council is a separate process that reviews from outside. Compound learning: failures become rules, rules fire on future tasks.
CREATE β WORK β MERGE β PAUSE. Nothing runs forever. Specialists merge back and pause. Idle agents retired after 3 pulses. The system cleans itself.
4 seats as loss functions, not personalities. Each optimizes a different objective β this prevents groupthink and ensures diverse pressure. Council thinks, CEO executes.
Every problem we identified in building this system, and the solution that's already built. This is why it can be a template.
Agent checks its own task queue instead of surveying the company. Paperclip's heartbeat prompt overrides custom instructions.
SOUL.md + AGENTS.md + HEARTBEAT.md + cron prompt + Paperclip instructions all say "YOU ARE NOT A WORKER." All 5 must be present or it falls back.
The brain has full terminal/file/web power β it CAN research and code. But that's a worker mindset. The brain should POINT, not BUILD.
"Hermes brain points, Paperclip hands execute." Pattern: identify root cause β encode in agent SOUL β wake the right agent β verify outcome.
Long conversations fill the context window. Agent loses early context, makes worse decisions.
Don't compress β UNPACK. Child inherits PURPOSE, not history. Max depth 2. Peer-to-peer graph, not tree. Lifecycle: CREATE β WORK β MERGE β PAUSE.
Supervisors that accumulate observations can consolidate them into misaligned self-models, even when told they're being tested (Anthropic, June 2025).
Do not audit yourself; be externally audited. The Seneschal does not score the Seneschal. The auditor is a different process. Stay Neo β enter any room, don't become it.
Multiple AI models from the same provider have Οβ0.7 error correlation. "Unanimous" isn't really independent.
4 different providers. Each seat optimizes a different objective (path length, false positives, information loss, local minima). Progressive dialectic: independent proposals before cross-critique.
"Fix the thing" β agent has no context β wastes entire run producing nothing useful.
Every brief includes: exact file path, specific line/function, expected output shape, acceptance criteria, model constraints. Brief depth = agent success rate.
Same work appears 3 times. CEO wastes every pulse re-reading duplicates.
Cross-reference backlog against done/in-progress. Cancel duplicates immediately with PATCH /api/issues/{id} {status: cancelled}.
At N > 1000, chi-square detects trivially small deviations. 0.5% difference produces p < 0.001. Causes false failures.
Gate on Cohen's w < 0.3, not p-value. Keep chi-square for diagnostics only. w < 0.1 = negligible, < 0.3 = acceptable, < 0.5 = investigate, β₯ 0.5 = reject.
Specialists created for one task never shut down. Costs accumulate.
Every child has time-to-live. Auto-pause after 30 min inactivity. Auto-destroy after 24h idle or task complete. Lifecycle is mandatory, not optional.
This isn't theory. Every role below is currently done by a human at a company. The ATLAS architecture replaces each with a persona that runs 24/7, costs cents per hour, never tires, and improves every iteration through compound learning.
| Human Role | What They Do | Agent Replacement | Cost Ratio |
|---|---|---|---|
| CEO / General Manager | Strategy, delegation, hiring, performance review | CEO Agent + High Council | ~1000:1 |
| Product Manager | PRDs, acceptance criteria, roadmap | Product Team Agent | ~500:1 |
| Software Developer | Code features, fix bugs, review PRs | Builder Agent + QA Agent | ~200:1 |
| Data Analyst | Find sources, validate, build pipelines | Scout + Validator Agents | ~300:1 |
| QA Engineer | Test suites, validation gates, regression | QA Agent (adversarial GAN) | ~400:1 |
| Researcher | Market research, competitor analysis, reports | Specialist (unpacked on demand) | ~500:1 |
| Compliance Officer | Regulatory checks, risk assessment | Compliance Specialist | ~300:1 |
| Strategy Consultant | Analysis, recommendations, frameworks | High Council (4 seats) | ~1000:1 |
| Project Coordinator | Task tracking, status reports, blockers | CEO Pulse (every 30 min) | ~800:1 |
Cost ratio = human annual salary : agent compute cost per year. Based on NZ market rates vs Ollama/local + cloud API usage. Agent costs include model inference, infrastructure, and API calls. Human costs include salary, overhead, training, management, and replacement.
The complete adaptive persona pipeline β the JSON that started this conversation, completed and fixed. This is the spec that powers the cognitive loop at every layer.
ATLAS β Adaptive Persona Agent Architecture
Built on Hermes Agent Β· Seneschal Pattern Β· CoALA Cognitive Architecture
Self-unpacking company Β· Russian Doll Protocol Β· High Council Governance
This is the template. Every layer is a persona. Every persona can spawn new personas. The recursion is the architecture.