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.
npx agentmods add agents/juspay/neurolink/artifact-bankinggit clone --depth 1 https://github.com/juspay/neurolinkWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/agents/juspay/neurolink/artifact-banking)<a href="https://agentmods.dev/agents/juspay/neurolink/artifact-banking"><img src="https://agentmods.dev/badge/agents/juspay/neurolink/artifact-banking.svg" alt="Measured on agentmods" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00000 | $0.02644 |
| Opus 5 | $0.00000 | $0.01322 |
| Sonnet 5 | $0.00000 | $0.00529 |
| Haiku 4.5 | $0.00000 | $0.00264 |
Grade A, and why
ARTIFACT-BANKING 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 today.
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.
How it starts
The opening of the file, as written. The whole thing — 239 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Artifact Banking (bankArtifact / readArtifact)
A long-running agent produces outputs that do not fit in a conversation: a worker's full report, a build log, a stage's structured result. The tempting answer is to truncate one and send the head — but the discarded bytes are gone, and nothing records that they ever existed.
Banking is the other answer. The payload is written to disk whole, and what goes into the conversation is a bounded preview plus the exact call that reads the rest. Context cost stays flat, evidence stays complete, and compaction can evict the preview without destroying anything.
Read-back is the tool that already existed: retrieve_context. No new tool,
no new name for the model to learn.
Quick start
import { NeuroLink } from "@juspay/neurolink";
const neurolink = new NeuroLink();
const ref = await neurolink.bankArtifact(fullWorkerReport, {
kind: "worker-report",
label: "delegate:auth-review",
sessionId: "review-1421",
});
// Put THIS in the conversation — never the report itself:
// ref.preview bounded head slice (1000 chars by default)
// ref.readBackHint the literal retrieve_context call that fetches the rest
// ref.sizeBytes how much is actually there
// And from host code, whenever you want it back:
const whole = await neurolink.readArtifact(ref.artifactId);
const window = await neurolink.readArtifact(ref.artifactId, {
offset: 100_000,
limit: 50_000,
});
The model reads the same artifact with the tool it already has:
// retrieve_context
{ "artifactId": "9f1c…", "offset": 0, "limit": 50000 }
// → { content, totalSize, hasMore, offset, limit }
retrieve_context is registered automatically the first time anything banks —
you do not have to configure mcp.outputLimits to get it.
API
| Member | Returns | Notes |
|---|---|---|
bankArtifact(payload, options) |
BankedArtifactRef |
Stores the payload whole; returns id + bounded preview |
readArtifact(id, page?) |
string | null |
Full payload when page is omitted; null if unknown |
getArtifactStore() |
ArtifactStore |
Creates the store on demand and registers retrieve_context |
setArtifactStore(store) |
void |
Swaps the backend for every path, MCP normalizer included |
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.
- today Changed · +126 lines 015e916e8df4
- 5d ago First seen · 113 lines · 0 tokens per session scan A 07cc5e3fec7a
ARTIFACT-BANKING is an agent published in the GitHub repository juspay/neurolink (128 stars, last pushed yesterday), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 2,644 tokens. 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.
Other agents, from other repositories
llm-integrator
LLM integration: OpenAI/Anthropic/Ollama APIs, prompt templates, function calling, streaming, token cost optimization.
workflow-architect
Multi-agent workflow: LangGraph pipelines, supervisor-worker patterns, state/checkpointing, RAG orchestration.
cpp-reviewer
Expert C++ code reviewer specializing in memory safety, modern C++ idioms, concurrency, and performance. Use for all C++ code changes. MUST BE USED for C++ projects.
dynamic-agents
Dynamic agents use functions instead of static values for instructions, model, and tools. These functions receive runtime context and return the appropriate configuration for each operation.
accessibility-specialist
Accessibility expert: WCAG 2.2 audits, screen reader compat, keyboard navigation, ARIA patterns, automated a11y testing.
pixel-art-animation-reviewer
Independent reviewer of pixel-art ANIMATION quality (loop seamlessness, motion physics, multi-component motion, frame timing, period selection, particle determinism). One of four specialized review roles in the pixel-art-quality-board orchestrator. Use when the user asks to "check animation timing", "verify loop…