Synthetic sample dashboard, not proof

184k repeated tokens become a 5.5k context pack.

This static artifact uses wholly synthetic, illustrative data, not a customer run or benchmark. It shows what a useful private dashboard should answer first: before and after tokens, estimated repeated context waste, claim boundaries, quality status, top repeated files, and what to run next.

Context estimate Usage-stretch Raw/unverified API meter Verified CostLayers attribution Quality gate
Run your local audit Open recipe Open private dashboard

Proof details

Small receipt, enough to act.

The public sample stays compact: no raw prompts, source files, provider responses, API keys, or private dashboard URLs.

Top repeated files

These are the files the sample agent kept rediscovering before the reusable context pack was generated.

Repeated filesrc/routes/checkout.tsRoute contracts and error handling reloaded across tasks.
48,300 tokens
Repeated filesrc/agent/context.tsAgent context builder and prompt assembly rules.
41,900 tokens
Repeated filedocs/architecture.mdStable architecture notes better referenced by hash.
33,600 tokens
Repeated fileREADME.mdSetup and workflow text repeatedly sent with every run.
22,800 tokens

Next commands

Start local, add quality evidence, then open a private dashboard. Use API mode only when you want invoice-backed provider-dollar proof.

# 0.8.61 immutable hosted build
# 1. Local context estimate
npx -y https://costlayers.com/costlayers-0.8.61.tgz audit --monthly-spend 1000

# 2. Quality gate / no-regression receipt
npx -y https://costlayers.com/costlayers-0.8.61.tgz quality compare -- npm test

# 3. Private dashboard URL
npx -y https://costlayers.com/costlayers-0.8.61.tgz dashboard

# 4. Privacy-safe proof bundle
npx -y https://costlayers.com/costlayers-0.8.61.tgz proof --refresh

0.8.61 is deployed at the immutable hosted URL and is not published to npm. Verify downloaded bytes against the hosted SHA-256 file.

Reproducible local demo

No separate public sample repo is attached. Use the local recipe and quarantine registry to verify public labels and immutable hashes without promoting raw legacy artifacts.

node benchmarks/world-best/claim-consistency-smoke.js
node benchmarks/world-best/quality-safe-proof-fixture-smoke.js
node -e 'const r=require("./proof_artifacts/legacy-artifact-registry.json"); console.log(r.registry_status, r.policy.promotion_status, r.artifacts.length)'

Expected labels include claim consistency smoke passed, quality-safe proof fixture smoke passed, active_quarantine, and prohibited. The consistency smoke recomputes every registered raw-file and screenshot hash.

Claim boundaries

Context estimateBefore/after tokens are local counters and hashes, useful for finding waste but not provider billing proof.
Usage-stretchNative Codex/ChatGPT mode can stretch usage limits; it does not claim lower flat subscription invoices.
API invoice proofInvoice-backed savings require routed API-mode traffic, gateway rows, and proof receipts for the same workflow.
Quality gateSavings count only when the chosen gate passes for direct and optimized paths on the same task class.
Old partial artifactIts mixed-arm raw/unverified meter delta does not isolate provider-native caching and is not verified CostLayers-attributed savings.

old_partial_artifact: legacy_verified_field_accepted=false; verified_costlayers_attributed_saved_cost_usd=$0.000000; provider_avoided=0; best-in-world=false