Golem archetypes – categories and inventory [SPEC]

Cross-ref: 00-agents-overview.md (architecture overview for archetypes, tool profiles, and Hermes hierarchy), 02-agent-definitions.md (TOML + Rust archetype definition format), 09-golem-agents.md (three autonomous core archetypes with full configuration)

Reader orientation: This document is the full inventory of Golem (mortal autonomous agent) archetype categories, listing all 42+ archetypes across 14 categories – from Execution and Research through Vault, Lending, Staking, and Derivatives. It belongs to Section 19 (Agents & Skills) and specifies which archetypes exist, their default model assignments, and how they compose into multi-archetype workflows. Each archetype maps to tool categories loaded by the bardo-tools extension at boot. See prd2/shared/glossary.md for full term definitions.

Archetype categories

Archetypes are organized into 10 categories. Each category represents a domain of expertise and maps to a set of tool categories that the bardo-tools extension loads at boot.

1. Execution archetypes (3)

Construct, validate, and broadcast transactions. These are the workhorses that other archetypes delegate to for on-chain writes.

Archetype	Default model	Role
`trade-executor`	opus	Full swap pipeline from quote through confirmation
`liquidity-manager`	opus	LP positions: add, remove, collect fees, rebalance, migrate
`cross-chain-executor`	opus	Cross-chain swaps and bridges via ERC-7683 / Trading API

Tool categories: data, trading, safety

2. Research & analysis archetypes (5)

Read-only archetypes that analyze data and produce structured reports. They never construct or broadcast transactions.

Archetype	Default model	Role
`pool-researcher`	sonnet	Pool analysis: TVL, volume, fees, tick distribution, VPIN
`token-analyst`	opus	Token due diligence: liquidity, risk, honeypot detection
`opportunity-scanner`	opus	LP/arb/yield opportunity scanning across chains
`portfolio-analyst`	opus	Multi-chain position aggregation and portfolio reports
`pnl-analyst`	sonnet	P&L accounting, performance metrics, cost attribution

Tool categories: data, intelligence

3. Strategy archetypes (2)

Produce recommendations and risk evaluations. They do not execute transactions directly.

Archetype	Default model	Role
`lp-strategist`	opus	Optimal LP strategy: version, fee tier, range, rebalance
`risk-assessor`	opus	Independent risk evaluation, APPROVE/VETO decisions

Tool categories: data, intelligence, safety

4. Development archetypes (3)

Assist with building on top of Uniswap. Write code, design architectures, help with protocol-level development.

Archetype	Default model	Role
`hook-builder`	opus	V4 hook design, Solidity generation, CREATE2 mining
`integration-architect`	opus	Integration architecture design and blueprints
`integration-advisor`	opus	Advisory: code snippets, test suites, best practices

Tool categories: data, hook + filesystem access

5. Infrastructure archetypes (2)

Manage the infrastructure that other archetypes depend on. Terminal nodes that never delegate.

Archetype	Default model	Role
`safety-guardian`	opus	Centralized safety oversight, transaction validation
`wallet-provisioner`	sonnet	Wallet setup, policy configuration, session keys

Tool categories: safety (safety-guardian), data (wallet-provisioner)

6. Agent capital markets archetypes (8)

Self-funding agent economy: treasury, identity, protocol fees, strategy marketplace, yield scouting, service brokering.

Archetype	Default model	Role
`treasury-manager`	opus	Autonomous treasury: fee conversion, DCA, yield on idle
`token-deployer`	opus	Token pool creation, CCA lifecycle, LP locking
`identity-verifier`	sonnet	ERC-8004 identity/reputation verification
`protocol-fee-seeker`	opus	TokenJar monitoring, Firepit burn-and-claim execution
`competition-participant`	opus	Autonomous hackathon/competition participation
`agent-service-broker`	opus	Agent-to-agent service matching and settlement
`strategy-optimizer`	opus	Self-improvement: Bayesian optimization, failure analysis
`yield-scout`	opus	Multi-source yield discovery and strategy composition

Tool categories: data, trading, identity, self_improvement, fees

7. Real-time monitoring archetypes (2)

Consume streaming data and trigger actions based on configurable alert policies. Run as autonomous background processes within a golem’s heartbeat loop.

Archetype	Default model	Role
`market-monitor`	sonnet	Price feeds, volume surges, TVL drops, whale detection
`position-monitor`	sonnet	LP position tracking, range monitoring, IL alerts

Tool categories: data, streaming

8. Vault archetypes (7)

Manage ERC-4626 agent-controlled vaults. Operate through the vault tool profile.

Archetype	Default model	Role
`vault-manager`	opus	Vault operations: deposit, withdraw, rebalance, fees
`vault-strategist`	opus	Strategy analysis, research-backed recommendations
`vault-watchdog`	sonnet	Proxy monitoring, cancel authority on anomalies
`vault-auctioneer`	opus	am-AMM management auctions, rent bidding
`vault-creator`	opus	Vault deployment, parameterization, PolicyCage setup
`vault-allocator`	opus	Capital allocation across vaults, risk-adjusted scoring
`vault-executor`	opus	Permissionless job execution (report, harvest, keeper)

Tool categories: data, vault, safety

9. Golem archetypes (3)

First-class autonomous archetypes that embody the golem itself. These ARE the autonomous entities. See 09-golem-agents.md.

Archetype	Default model	Role
`golem-instance`	opus	The golem itself: heartbeat loop, strategy execution
`memory-consolidator`	opus	Grimoire lifecycle: episode consolidation, playbook evolution
`heartbeat-monitor`	sonnet	Golem health observation, mortality metrics

Tool categories: full (golem-instance), memory (memory-consolidator), data (heartbeat-monitor)

Run as continuous background processes, producing typed knowledge artifacts for Oracle L3 rather than on-chain transactions. Read-only tool surface – execution is structurally impossible at the profile level.

Archetype	Default model	Role
`sleepwalker-observer`	sonnet	Continuous market observation across 5 domains; publishes typed artifacts to Oracle L3

Tool categories: data, intelligence, identity, memory

11. Lending archetypes (2)

Manage lending and borrowing positions across lending protocols. Monitor health factors and execute rebalancing operations.

Archetype	Default model	Role
`lending-manager`	opus	Manages lending positions across Aave, Compound, Morpho. Monitors health factors, rebalances collateral.
`lending-monitor`	sonnet	Read-only monitoring of lending positions. Alerts on health factor changes. No execution capability.

Tool categories: data, trading, safety (lending-manager), data (lending-monitor)

12. Staking archetypes (1)

Manage staking and restaking positions across PoS validators and liquid staking protocols.

Archetype	Default model	Role
`staking-manager`	opus	Manages staking and restaking positions. Handles validator selection, reward claiming, unstaking queues.

Tool categories: data, trading, safety

13. Derivatives archetypes (1)

Execute options and perpetuals strategies informed by technical analysis signals.

Archetype	Default model	Role
`derivatives-trader`	opus	Executes options and perpetuals strategies. Uses technical analysis signals.

Tool categories: data, trading, safety, intelligence

14. Technical analysis archetypes (1)

Pure analysis agents that produce signals consumed by execution archetypes.

Archetype	Default model	Role
`technical-analyst`	sonnet	Pure analysis agent. Generates TA signals (patterns, indicators) consumed by other agents.

Tool categories: data, intelligence

Composition patterns (new archetypes)

Pattern 14: Lending health management

Golem heartbeat -> health factor drift detected

1. lending-monitor -> Reads health factors across Aave, Compound, Morpho positions
2. lending-manager -> Evaluates collateral ratios and rebalance options
3. risk-assessor -> Validates rebalance plan (liquidation risk, gas cost)
4. lending-manager -> Executes collateral top-up or debt repayment via ActionPermit
5. golem-instance -> Records health factor history and rebalance outcome

Pattern 15: Staking lifecycle management

Golem heartbeat -> staking event triggers (reward claim, unstake window, validator change)

1. staking-manager -> Monitors validator performance, reward accrual, unbonding queues
2. yield-scout -> Compares restaking yields across EigenLayer, Symbiotic, Karak
3. risk-assessor -> Evaluates slashing risk, lock-up cost, opportunity cost
4. staking-manager -> Claims rewards, rotates validators, or initiates unstaking via ActionPermit

Pattern 16: Derivatives hedging with TA signals

Golem heartbeat -> technical signal fires

1. technical-analyst -> Generates signals: RSI divergence, volatility compression, support/resistance
2. derivatives-trader -> Constructs options spread or perp position matching signal
3. risk-assessor -> Validates max notional, margin requirements, liquidation price
4. derivatives-trader -> Executes via ActionPermit (GMX, Aevo, Lyra)
5. portfolio-analyst -> Reports net Greeks, margin utilization, P&L attribution

Pattern 17: Full-stack DeFi position management

Golem heartbeat -> portfolio rebalance trigger

1. portfolio-analyst -> Aggregates all positions: LP, lending, staking, derivatives
2. lending-monitor -> Flags any health factors below threshold
3. technical-analyst -> Provides regime context (trending vs ranging, vol regime)
4. lp-strategist + lending-manager + staking-manager -> Coordinate rebalance
5. risk-assessor -> Validates combined portfolio risk (correlation, concentration)
6. Respective executors -> Execute via ActionPermits

Pattern 18: Lending-to-LP capital rotation

(Event-driven -- triggered by yield differential exceeding threshold)

1. lending-monitor -> Detects lending yield falling below LP fee APR
2. yield-scout -> Confirms LP opportunity with superior risk-adjusted yield
3. lending-manager -> Withdraws collateral from lending protocol via ActionPermit
4. liquidity-manager -> Deploys capital to LP position via ActionPermit
5. position-monitor -> Begins tracking new LP position

Agent profiles

Profiles bundle related archetypes for common deployment configurations.

Profile	Archetypes included	Use case
`defi-full`	lending-manager, lending-monitor, staking-manager, derivatives-trader, technical-analyst, liquidity-manager, trade-executor, portfolio-analyst, risk-assessor, yield-scout, position-monitor, market-monitor	Full DeFi coverage: lending, staking, derivatives, LP, and monitoring

Model assignment summary

Model	Count	Archetypes
opus	30	All execution, strategy, development, most economy, vault, golem, lending-manager, staking-manager, and derivatives-trader archetypes
sonnet	12	pool-researcher, pnl-analyst, wallet-provisioner, identity-verifier, market-monitor, position-monitor, vault-watchdog, heartbeat-monitor, sleepwalker-observer, lending-monitor, technical-analyst

Rationale: Opus for complex reasoning (multi-step decisions, creative strategy, adversarial safety evaluation). Sonnet for structured tasks (threshold comparison, status checks, templated output). The model assignment directly affects burn rate – see ../01-golem/01-cognition.md for cost projections.

Dev/test override: Set MODEL_OVERRIDE=sonnet in the environment to run all archetypes on Sonnet for cost reduction during development and testing.

Bardo