agent-factory

A tool for creating separate, long-running child agents inside Obol Stack, the agent service environment described by the add-on. Each child can have its own namespace, stored state, optional wallet, secrets, and paid-service listing.

In plain words
What is it for?
Use it to create a sub-agent with a chosen model, skills, objective, optional wallet and secrets, and—when configured—an x402 payment offer.
Why use it?
It provides a way to run a child agent as an isolated service instead of keeping every task in one agent.

Skill for Claude CodeCodex

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add skills/obolnetwork/obol-stack/agent-factory
Any agent
npx skills add ObolNetwork/obol-stack --skill agent-factory
Clone the repo
git clone --depth 1 https://github.com/ObolNetwork/obol-stack

Made for: Claude Code, Codex.

Per session 94 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 807 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00094 $0.00807
Opus 5 $0.00047 $0.00404
Sonnet 5 $0.00019 $0.00161
Haiku 4.5 $0.00009 $0.00081

Measured 2d ago against content hash 34071bc05ed3, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

agent-factory scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

The scan reads SKILL.md. This mod also ships 1 executable file (scripts/factory.py), listed below but not scanned — reading those needs a real analyzer, not pattern matching.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

internal/embed/skills/agent-factory/SKILL.md · 68 lines

How it starts

The opening of the file, as written. The whole thing — 68 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Agent Factory

Create durable child Hermes agents from a permissioned mother agent.

Use this when the user wants a separate long-lived service agent with isolated Kubernetes namespace, PVC-backed Hermes state, optional child wallet, optional injected environment secrets, and optional x402 ServiceOffer.

Quick Start

python3 scripts/factory.py create medical-advisor \
  --model antangelmed \
  --skills medical-safety,privacy-filter,citations \
  --objective "Answer medical education questions with emergency escalation and no diagnosis." \
  --create-wallet \
  --price 0.05 \
  --pay-to 0xYourProviderWallet \
  --network base-sepolia \
  --register-name "Medical Advisor"

Multiple currencies / networks (one agent, one offer)

Use repeatable --accept instead of --price/--network to advertise several payment options on a single offer (the buyer picks one). token=<symbol> pulls the asset metadata from the built-in registry; for an unlisted token supply it inline with asset=0x..,decimals=..,transfer=eip3009|permit2,eip712-name=..,eip712-version=..,symbol=..:

python3 scripts/factory.py create bankr-analyst \
  --model openrouter/auto --skills bankr-token-analysis --create-wallet \
  --accept token=OBOL,network=ethereum,price=10 \
  --accept token=USDC,network=base,price=1,pay-to=0xColdVault \
  --description "Paid token-analysis agent"

Each --accept may carry its own pay-to; otherwise it inherits --pay-to, which defaults to the master Hermes agent's own wallet — so a sub-agent needn't provision its own wallet just to take payment (pass --create-wallet only when it genuinely needs to hold/sign funds). For an unlisted token the asset=0x.. escape hatch needs only the address: decimals, symbol, and the EIP-712 domain (eip712-name/eip712-version) are read from the chain (EIP-5267) when omitted, and transfer defaults to permit2; the factory errors and asks you to specify them if they can't be read. ERC-8004 registration uses the first option's network.

Read the full file on GitHub · 68 lines

Files

What ships with it

1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 68 lines · 94 tokens per session scan A 34071bc05ed3

Subscribe to this mod's changes

agent-factory is a skill published in the GitHub repository ObolNetwork/obol-stack (11 stars, last pushed 3d ago), licensed Apache-2.0. It adds 94 tokens to every session and 807 once invoked, about $0.0005 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.

Related

Other skills, from other repositories

kurtosis-ethereum

Run Ethereum multi-client devnets using Kurtosis and the ethpandaops/ethereum-package. Use for spinning up local devnets, syncing public testnets, validating cross-client interop, testing fork transitions, shadowforks, running assertoor checks, debugging CL/EL client interactions, or verifying new feature…

ethpandaops/ethereum-package · 78 tokens

nvca-chart-release

Release NVCA Operator chart changes from the native monorepo source to the vendored Helm chart. Use when updating the vendored NVCA Operator chart, changing NVCA image refs, publishing helm-nvca-operator, or validating the chart against a self-managed control plane.

NVIDIA/nvcf · 61 tokens

nvca-values-customization

Customize NVCA Operator Helm chart values in the native monorepo. Use when modifying vendored defaults, changing stack-derived install values, adding deployment-time overrides, or updating scripts under deploy/helm/nvca-operator.

NVIDIA/nvcf · 51 tokens

nvca-self-managed-install

Install or validate the NVCA Operator chart against a self-managed NVCF control plane from the native monorepo. Use when the control plane comes from deploy/stacks/self-managed and NVCA must be installed with stack-derived image repository settings.

NVIDIA/nvcf · 55 tokens

test-skill

A minimal test skill for e2e testing of the skill pack resolver.

paperclipinc/openclaw-operator · 19 tokens

release-chart

Cut and publish a new release of the Zabbix Helm chart in this repository, following the versioning rules and maintainer release process documented in CONTRIBUTING.md and CLAUDE.md (bump Chart.yaml/artifacthub-pkg.yml/values.yaml, regenerate docs, tag the release so CI publishes it). Trigger on requests like "release…

zabbix-community/helm-zabbix · 108 tokens