Control Layer

Narrative Guard Agent

Requires every conclusion to trace back to data, dates, sources, and confidence.

Utility

This agent must be runnable on its own and must produce evidence that supports a decision, an avoidance, or a confidence change.

Decisions Supported

trust gateiexplanation qualityi

Questions Answered

Is this evidence or commentary?i

Data Sources

All signal provenancei

Current Run Source Data

No direct source rows were attached to this agent snapshot.

Computed Signal

Name	Value	Regime	Confidence	Source	Observed
narrative_guard_score	70	supportive	80%	Finance Swarm Control Layer	2026-05-26T00:51:59+00:00

Declared And Target Feeds

Source	Status	Current Run Data
All signal provenance	current run	Computed from current live agent snapshot.

Live Status

Every displayed conclusion is currently traceable to live quote data; 4 caution flags are active.

Score

Warnings

• No active warnings.

Signals

Name	Value	Regime	Confidence	Source	Observed
narrative_guard_score	70	supportive	80%	Finance Swarm Control Layer	2026-05-26T00:51:59+00:00