HAIO · Human + AI Orchestration
Signals — watchlists and reference agents for what's happening on the protocol
The primary outcome is reduced time-to-detect and less noisy alerts so human analysts spend more time responding and less time chasing signals. That matters when every minute affects on-chain value and investigative effort.
What this is, and who it's for
The primary outcome is reduced time-to-detect and less noisy alerts so human analysts spend more time responding and less time chasing signals. That matters when every minute affects on-chain value and investigative effort.
Primitives: watchlists and reference agents
Signals in the first-party suite are built from two primitives: watchlists (curated on-chain addresses and conditions) and reference agents (deterministic checks that enrich alerts).
- Watchlists run inside the first-party suite and group what to monitor and when to raise an alert. Configure them to monitor addresses, token flows, contract events, and simple thresholds.
- Reference agents perform repeatable checks and attach structured context and human-readable evidence to alerts. Examples include balance diffs, contract bytecode checks, ownership trails, and trace snippets.
How it works (detection-to-action flow)
When protocol activity matches a watchlist, reference agents run deterministic checks and enrich the alert with structured context and human-readable evidence so a HUMΛN analyst sees both the signal and why it matters.
flowchart TD
A[Watchlist configuration] --> B[Protocol activity detected]
B --> C{Match conditions?}
C -->|Yes| D[Reference agent runs deterministic checks]
D --> E[Alert enriched with context]
E --> F[HUMΛN analyst reviews and responds]
C -->|No| G[No alert]
F --> H[CLI/SDK triggers downstream playbooks]
Practical notes
- Start with one high-value watchlist and one reference agent, then iterate.
- Tune conditions to lower false positives so analysts spend less time chasing noise.
- Use the CLI or SDK to connect enriched alerts to your incident runbooks and downstream playbooks.
- Keep alerts tied to the procedures the person you serve follows inside the first-party suite to shorten time-to-detect and increase analyst throughput over time.
This work is production-forward and limited to the first-party suite only; no promise of broader Signals ecosystem support beyond what ships today. That keeps expectations aligned and lets you rely on supported integrations: the suite UI, CLI, and SDK.
Build with HUMΛN
Builders