Audit configured evidence, watched-market records, prices, policy, alerts, paper orders, wallet snapshots, journals, analytics, and exports in TEMIRIN.
August 3, 2026 · 4 min read
A prediction-market evidence audit begins with the records TEMIRIN currently stores: configured source items, watchlist-derived market records, trigger and current prices, public-wallet snapshots, policy state, alert attempts, authenticated decisions, user-entered paper orders, and journal notes. Select a workspace and date range before reviewing totals. Stable identifiers and labeled timestamps let another authorized teammate reopen the same material and distinguish each record type.
Use the exact market ID, question, outcome, close time, and resolution source when reconciling records. Evidence appears beside a watched market through its explicit market identifier. Keep unlinked source items in their source context and record missing fields as unavailable. This creates a reproducible scope for the audit and prevents a market title, screenshot, or exported row from becoming detached from the contract it describes.
For RSS and HTTPS records, inspect the configured source, canonical URL, observed time, available publication time, freshness, reliability, relevance, direction, and duplicate state. Provider-configured Telegram and X readers retain their own source references where available. A deterministic brief organizes these stored fields so the reviewer can move from summary to source and record an attributable interpretation.
Sample both fresh and stale records, repeated items, corrections, and collection failures. Confirm that each brief uses the expected explicit market ID and that the source trail remains accessible. When evidence changes, keep the later source record and journal note attributable. The audit should explain what the workspace stored at each time rather than flattening several observations into one undated conclusion.
Every sampled opportunity should retain its first trigger price and a separately timestamped current price. Where available, inspect public bid, ask, spread, and visible depth with the exact market and outcome identity. These observations show the market context a reviewer saw at a point in time and support consistent journal analysis across watchlist records.
Check that later synchronization updates current context while preserving the trigger reference. Compare paper entry and mark fields only within the paper ledger, and keep trigger-price comparisons labeled separately. This distinction makes exported analysis easier to reproduce and prevents a later market observation from silently replacing the starting point used by the original review.
A configured public wallet supplies read-only snapshots, positions, PnL context, and concentration with its own observation time. Keep these records separate from paper orders and document incomplete cost basis or market coverage. The snapshot provides portfolio context for review while retaining its identity as public external data.
For the same sample, inspect the stored stake cap, source and category enabled state, applicable maximum-stake overrides, free-capital percentage, price-move guard, and kill switch. Preserve each pass or block reason. Policy history explains the workspace boundary that applied to a review or paper request and should remain distinguishable from the evidence and wallet records considered by the user.
Alert rules create in-app events and provider-configured email, Telegram, Slack, or outbound webhook attempts. Inspect destination, attempt time, status, safe diagnostics, and manual retry history. A later successful attempt should remain attached to the original alert so the audit can show transport recovery without losing the earlier failure.
Authenticated approve or reject events carry their own actor and time. User-entered paper orders then receive per-order, rolling twenty-four-hour, per-market exposure, and portfolio-concentration checks in a separate ledger. Trace shared opportunity identifiers across these records while preserving their distinct statuses. This makes the review sequence understandable without combining transport, decision, and simulation outcomes.
Journal analytics and export support a repeatable closing review. Export the selected record types and include stable IDs, market identity, labeled timestamps, relevant status, and paper-only markers. Exclude credentials and keep source, wallet, decision, alert, and paper columns clearly named. Another authorized teammate should be able to select the same period and reconcile the exported sample with protected workspace records.
Finish with an exceptions list covering stale sources, unlinked evidence, missing prices, incomplete wallet snapshots, policy blocks, failed deliveries, and rejected paper requests. Assign an owner where a current configuration or operating note needs attention. Append later findings as new journal context so the original audit sample remains reproducible and useful for team review.
Configured evidence, deterministic briefs, watchlist-derived market records, prices, public-wallet snapshots, policy state, alert attempts, in-app decisions, paper orders, journals, analytics, and exports.
Use a defined workspace and date range, retain stable record IDs and labeled timestamps, and reconcile a sample back to the authenticated workspace.
Keep them separate because public-wallet snapshots and user-entered paper simulations have different provenance and timestamps.
