CRO PlatformAI-Native CRO OS
Toggle navigationMenu

Feature Flags

Every flag the platform recognises — agent toggles, UI rollouts, integration kill switches, and the regulated controls that gate Part 11 promotion.

Read-only view. Toggle in Vercel/Supabase env config.

Total

22

On

12

Off

3

Staged

3

Rollout

4

Agents

4 flags
NameDescriptionStateTargetOwnerLast toggled

agent-30-rag-enabled

Agent 30 — Knowledge RAG

Enables the knowledge RAG agent to answer cross-study questions with cited passages from the vector spine.onplatform-eng27d ago

agent-27-audit-trail-enabled

Agent 27 — Audit Trail

Continuously synthesizes inspection-grade audit narratives from the provenance event spine.oncompliance1mo ago

agent-04-essential-docs-enabled

Agent 04 — Essential Documents

Drafts and files ICH-E6 essential documents into the TMF zone, gated by Part 11 signature.ontmf-ops1mo ago

agent-31-evaluation-enabled

Agent 31 — Evaluation

depends on: agent-30-rag-enabled

Runs the LLM-judge evaluator across staged outputs from authoring agents before any signature gate.rollout

50%

platform-eng17d ago

Regulatory

4 flags
NameDescriptionStateTargetOwnerLast toggled

staging-promotion-dual-control

Dual-control promotion

Require two distinct signed reviewers before any staging→live promotion of regulated records.stagedcompliance20d ago

inspection-zip-export

Inspection ZIP export

Generates a downloadable evidence ZIP from the inspection package, signed + checksum-verified.oncompliance1mo ago

auth-totp-required

TOTP required for signing

Requires a TOTP second factor on every Part 11 signature, even for re-auth within the session.onsecurity3mo ago

regulatory-ectd-publish

eCTD publishing

Generates eCTD-compliant submission backbone XML for module 5 study reports; requires regulatory sign-off.offregulatory5mo ago

Integrations

7 flags
NameDescriptionStateTargetOwnerLast toggled

anthropic-prompt-caching

Anthropic prompt caching

Use Anthropic prompt caching for long system prompts shared across agent invocations.onplatform-eng1mo ago

crons-paused

Crons paused (pause-by-default)

When on, every scheduled cron short-circuits without invoking Anthropic. Manual /admin/agents triggers bypass the pause. Driven by the CRONS_PAUSED env var; unset is treated as on.onplatform-eng11d ago

cost-cap-enforcement

Per-study cost cap

Enforces a hard per-study daily AI spend cap; agents above cap are paused until next window.onfinops2mo ago

notifications-email-delivery

Email notification delivery

Sends digest and signature-request emails through the notifier service; off uses in-app only.onplatform-eng3mo ago

provenance-supabase-realtime

Provenance Realtime channel

Publishes every provenance_event insert to a Supabase Realtime channel for live dashboards.onplatform-eng1mo ago

integration-medidata-rave

Medidata Rave integration

Bidirectional sync with Medidata Rave EDC for query reconciliation and SDV evidence.stagedintegrations1mo ago

integration-veeva-vault-etmf

Veeva Vault eTMF integration

Filing-side integration with Veeva Vault eTMF for essential document storage.onintegrations1mo ago

UI

5 flags
NameDescriptionStateTargetOwnerLast toggled

risk-dashboard-realtime

Risk dashboard realtime

depends on: provenance-supabase-realtime

Stream new risk register entries to the /risks dashboard via Supabase Realtime.rollout

75%

risk-mgmt24d ago

sponsor-portal-public-access

Sponsor portal public access

Allows magic-link-only access to the sponsor portal without Supabase Auth account creation.offsponsor-ops3mo ago

ui-new-global-nav

New global navigation

Refactored grouped global nav with section headings. Rolling out behind a flag for visual QA.ondesign-eng1mo ago

ui-roadmap-quarterly-view

Roadmap quarterly view

Adds a quarterly grouping toggle to the /roadmap page.rollout

25%

design-eng22d ago

ui-inspection-pdf-preview

Inspection PDF preview

In-browser PDF preview for evidence binder artifacts on the /inspection page.rollout

60%

design-eng16d ago

Experimental

2 flags
NameDescriptionStateTargetOwnerLast toggled

experimental-evaluator-double-check

Evaluator double-check (experimental)

depends on: agent-31-evaluation-enabled

Runs a second evaluator pass with a different model when the first pass returns borderline scores.stagedresearch18d ago

experimental-graph-reasoning

Graph reasoning (experimental)

Spike: optional graph-traversal step for cross-study reasoning. Off pending ADR re-evaluation.offresearch3mo ago