Governed AI Workforce · SOC Lead

The SOC Lead agent, under zero-trust governance.

A frontier-capable incident synthesizer that reads two upstream analyst feeds, decides severity, and escalates to opus only when the incident earns it — every call minted, policy-checked, and signed into an audit chain.

Asphodel + Tiresias-ZT · ACME Corporation
The Stakes

An autonomous SOC role is exactly where "trust the model" breaks.

The SOC Lead ingests threat intelligence and analyst findings, then reaches for the most expensive frontier model on the roster the moment it smells a breach. Ungoverned, that is an unbounded blast radius: unbounded spend, unbounded data reach, and no defensible record of why any call was made.

Without governance

Model picks its own tier. Reads any tenant's memory. No signed evidence of the decision. A prompt-injected "critical" burns frontier budget on demand.

Under Tiresias-ZT

Deny-by-default policy at the PEP. Per-agent capability token. RLS-isolated data tenant. Every hand-off and escalation written to an Ed25519 audit chain.

The Workforce

The SOC Lead is one governed identity inside a fleet of 367.

This is a live instance — 367 agents registered as control-plane tenants, each mapped to a real principal in the ACME directory and run under the same zero-trust policy the SOC Lead is bound by.

367
governed agents
7
showcase hand-off agents
360
fleet agents
325,154
directory principals
Source: registry (367 agent tenants) · personas.json (7) · agent-manifest.json (360) · tzt_principal — 28 departments, 10 regions, 626 roles.
Architecture Slice

The SOC Lead sits at a junction in a 14-edge hand-off DAG.

Two upstream agents feed it. It synthesizes, decides, and passes a hardened brief downstream. Nothing about that flow is implicit — each edge is a governed call through the enforcement plane.

threat-intel-researcher security-analyst soc-lead research-lead data-analyst content-editor compliance-officer
upstreamTwo feeds convergethreat-intel-researcher findings + security-analyst triage — the SOC Lead's only inputs2 upstreams
tiersHome / ceiling / floor per agentkeyword escalation (+incident / +breach / +critical) lifts toward the ceilinghome·ceiling·floor
degradeFallback-to-floor on LLMErrora provider failure never stalls the SOC pipelinesafe-degrade
Source: personas.json · runner.py schedule — 7 showcase agents, 14-edge topological hand-off DAG.
Live Proof

Every SOC Lead call takes the same governed path — no exceptions.

The runner mints a per-agent HS256 capability token from its 64-hex secret. From there the request is validated, policy-evaluated deny-by-default, and forwarded verbatim to the agent's own data tenant.

runner mints cap-token PEP :8343 — validate + policy eval soul-svc tenant PostgREST acme_soul (RLS write)

On allow, the PEP signs the call into its audit chain and ships it to the control plane. Usage lands in tzt_usage and surfaces on the Grafana ROI board.

Source: governed request flow — PEP host :8343, deny-by-default, forwards verbatim; audit chain → CP; usage → /cp/usage/report → tzt_usage → Grafana.
Walkthrough · 1 of 3

Synthesis: two upstream outputs become one governed judgment.

The SOC Lead does not gather its own intelligence. It reads what the threat-intel researcher and the security analyst already produced — both written to their own RLS-isolated tenants — and fuses them into a single severity call.

Input A — threat-intel-researcher

Upstream domain brief on emerging indicators, delivered as a governed CoT + memory write in its own soul-svc tenant.

Input B — security-analyst

Triaged findings from the analyst stage — the second and final upstream edge feeding the SOC Lead node.

The SOC Lead reads across the hand-off — but only what policy grants. Cross-tenant reach requires an explicit service role; it is never the default.
Source: personas.json hand-off DAG (2 upstream edges into soc-lead) · Plane 4 — FORCE RLS, cross-tenant needs SET ROLE service_role.
Walkthrough · 2 of 3

Escalation: "critical" is a governed decision, not a free reach for opus.

Keyword escalation is the mechanism — when the synthesized incident carries +incident, +breach, or +critical, the agent lifts from its home tier toward its ceiling. The frontier model is reachable, but bounded by the ceiling policy.

triggerKeyword escalation+incident / +breach / +critical in the synthesized briefrule-based
ceilingFrontier tier — opus-4.8reached only for high-severity, and only up to the agent's ceilingopus-4.8
homeRoutine severity stays on cheaper tierssonnet-4.5 and free/Ollama tiers carry the everyday loadsonnet · free
floorFallback-to-floor on LLMErrora degraded escalation still completes rather than failing opensafe-degrade
Source: personas.json — per-agent home/ceiling/floor tiers + keyword escalation, fallback-to-floor on LLMError · config.json roster (opus-4.8, sonnet-4.5).
Walkthrough · 3 of 3

The hand-off: a signed, isolated write — then downstream to research-lead.

The SOC Lead's judgment is not a chat message. It is a governed CoT and memory write into its own tenant, validated end to end, and it becomes the input to the next stage of the DAG.

mintRunner mints per-agent HS256 capability tokenfrom the agent's 64-hex capability secretHS256
enforcePEP validates api-key + cap-token, evaluates policydeny-by-default; on allow, forwarded verbatim:8343
writeCoT capture + memory write to acme_soulRLS-forced, tenant-isolated — POST /v1/cot/capture, /v1/memory/writeRLS
signPEP signs the call into its audit chainSoulKey Ed25519, durable WAL at /data/audit-wal, ships to CPEd25519
passOutput flows downstream to research-leadthe next of the 14 edges in the hand-off DAGnext edge
Source: Plane 3/4/5 — governed request flow · signed audit chain (SoulKey Ed25519) · endpoints /v1/cot/capture, /v1/memory/write.
Governance & Audit

Every decision is reviewable, and drift is contained automatically.

The SOC Lead operates under continuous behavioral monitoring. If its trajectory drifts, SoulWatch quarantines it without waiting for a human — and an analyst can verify the whole audit chain after the fact.

Behavioral containment

SoulWatch scores behavioral drift and auto-quarantines the agent into a shared postgres store. RFC5424 syslog streams to Promtail:1514 → Loki as SIEM.

After-the-fact review

The PEP /admin/* console exposes audit/verify, soulwatch/status, quarantine, and analytics — trajectory, model_baseline, campaign, drift. Role-token gated at pdp.tiresias.watch.

Source: Plane 3 SoulWatch → auto-quarantine, SIEM RFC5424 → Promtail:1514 → Loki · Review console = PEP /admin/* (pdp.tiresias.watch).
Economics

Governing the tier is governing the bill.

Because escalation is bounded by policy, opus is reserved for the incidents that warrant it. The observed per-call economics show why that boundary matters — the frontier tier costs two orders of magnitude more than the floor.

Model tierRole for SOC LeadObserved $/call
opus-4.8Ceiling — high-severity only$0.19
sonnet-4.5Routine synthesis$0.03
deepseek-v4-proCheap capable tier$0.002
gpt-oss:120bSelf-host tier$0.0015
gemma:freeFree floor$0.00
Source: ROI dashboard uid acme-roi-economics — observed per-call cost from tzt_usage.
Close

A frontier-capable SOC Lead you can actually put in production.

Two upstreams synthesized, severity decided, opus reached only when it is earned — and every step minted, policy-checked, RLS-isolated, and signed. The capability is the easy part. The governance is what makes it deployable.

2
upstream feeds synthesized
deny
default policy posture
Ed25519
signed audit chain
$0.19
bounded frontier call
Asphodel + Tiresias-ZT · ACME Corporation · governed, live, on GCP
1 / 1