ACS Agent Workflow
DraftStatus: Draft Version: 0.1.0 Last Updated: 2026-05-17 Owner: Axodus BBA / ACS
Human-Controlled Workflow
A BBA owner submits a bounded task such as research organization, claim screening, audience analysis, or draft preparation. ACS may return structured findings, missing-evidence flags, risk classifications, and suggested edits. The responsible human reviews sources, rejects unsupported output, and decides whether the material can advance.
ACS does not approve campaigns, publish content, contact users, purchase media, or certify legal compliance. Sensitive claims route to the appropriate product, legal, security, governance, or treasury reviewer. If the evidence is missing or contradictory, the workflow ends in revision or rejection, not automated publication.
Purpose
Define how ACS supports BBA work without replacing human accountability.
Scope
ACS may support research, drafting, campaign analysis, claim review, risk detection, report drafting, governance briefs, and narrative coherence.
Agent Roles
Morpheus supports narrative, positioning, governance alignment, and ecosystem coherence. Agent Smith supports adversarial review, hype detection, unsafe claim detection, and risk stress testing. Trinity supports Trading, Treasury, PnL, and performance claim review.
Workflow
ACS output should be generated from repository context, reviewed by the appropriate human or domain owner, escalated when sensitive, and recorded when used in public materials.
Governance Touchpoints
ACS may prepare governance communication drafts, but it cannot approve claims, campaigns, proposals, or public launch status.
Risk Considerations
ACS may be wrong. Sensitive claims require human validation and domain review before publication.