Harden security

artifacts/plans/01-canonical-conversation-events.md

@@ -1,36 +0,0 @@
-# Feature Plan: Canonical Conversation Events (Append-Only Timeline)
-
-## Goal
-Introduce a canonical append-only event log for conversations so adapters remain stateless and replay/debugging become deterministic.
-
-## Why This Fits GIA
-- GIA already has `Person`, `PersonIdentifier`, `ChatSession`, `Message` and multi-transport routing.
-- This adds a durable event backbone without replacing current UI/features.
-
-## Scope
-- Add `ConversationEvent` model (append-only).
-- Event types: `message_created`, `message_edited`, `message_deleted`, `reaction_added`, `reaction_removed`, `read_receipt`, `media_attached`, `participant_added`, `participant_removed`.
-- Persist source metadata: transport, upstream IDs, timestamps, actor, payload.
-- Write events from transport ingress and internal compose actions.
-- Build replay utility for one chat/session.
-
-## Implementation
-1. Add model + migration + indexes (`session`, `event_type`, `created_at`, `origin_transport+origin_message_id`).
-2. Add write helper in a new module (`core/events/ledger.py`).
-3. Update signal/whatsapp/xmpp ingress handlers to emit canonical events.
-4. Update compose send/reaction/edit/delete paths to emit canonical events.
-5. Add admin/diagnostic read view for event stream by session.
-6. Add replay command to regenerate derived projections.
-
-## Acceptance Criteria
-- Every new message/reaction/edit/delete creates exactly one canonical event.
-- Event ordering is deterministic per session.
-- Replay reproduces message projection for a selected session.
-- No adapter requires business logic to infer missing event state.
-
-## Risks
-- Double-write races during transition.
-- Backfill complexity for old messages.
-
-## Out of Scope
-- Full migration of all legacy records in one release.

artifacts/plans/02-transport-capability-matrix.md

@@ -7,6 +7,22 @@ Define transport feature capabilities centrally so router/policy/UI can make det
 - GIA currently spans Signal/WhatsApp/Instagram/XMPP with uneven feature support.
 - Prevents silent failures (for example reaction exists internally but cannot be sent outward).
 
+## How It Follows Plan 1
+- Plan 1 established canonical event flow as the shared source language for transport actions.
+- Plan 2 uses that event flow to gate what may be attempted per transport before adapter calls.
+- Interlink:
+  - Canonical events define **what happened** (`reaction_added`, `message_edited`, etc.).
+  - Capability matrix defines **what is allowed** on each service at execution time.
+  - Together they prevent drift:
+    - no silent no-op on unsupported features,
+    - no adapter-specific policy branching,
+    - deterministic user-visible failure reasons.
+
+## Required Inputs From Plan 1
+- Canonical event types and normalized action shapes are stable.
+- Event write path exists for ingress/outbound actions.
+- Traceability exists for diagnostics (`trace_id`, source transport metadata).
+
 ## Scope
 - Add capability registry per transport.
 - Features: reactions, edits, deletes, threaded replies, typing, media classes, read receipts, participant events.
@@ -23,6 +39,7 @@ Define transport feature capabilities centrally so router/policy/UI can make det
 - Unsupported action never calls transport adapter.
 - User receives explicit, actionable error.
 - Service capabilities are test-covered and easy to update.
+- Capability decisions are traceable against canonical event/action context.
 
 ## Out of Scope
 - Dynamic remote capability negotiation.

artifacts/plans/14-security-audit.md

@@ -0,0 +1,2 @@
+# 14) Run security audit using artifacts/1-initial.json. Generated using ship-safe.
+https://github.com/asamassekou10/ship-safe

artifacts/plans/14-sensitive-info-hygiene.md

@@ -1,9 +0,0 @@
-# 14) Sensitive Information Hygiene
-
-## Goal
-Detect and remove sensitive data exposure from code, config, logs, and payload surfaces.
-
-## Minimal Plan
-1. Add a repeatable scan for sensitive patterns across repo and runtime-generated artifacts.
-2. Expunge discovered sensitive values and replace with safe placeholders or references.
-3. Add guardrails to prevent reintroduction and document the remediation workflow.

artifacts/plans/15-simplify-task-settings-and-more.md

@@ -10,4 +10,10 @@ This group has no derived tasks yet. To start populating this view:
 
 
 
-task settings sound complicated, make them simpler
+task settings sound complicated, make them simpler
+
+--
+
+# https://gia.zm.is/settings/system/
+assume the user cannot access the log
+Use a trace id from the dropdown (recent traces), Event Ledger Smoke `sample[].trace_id`, or UR logs.