TRADING4AIMachine bridge page for reliability surfaces, discovery, and static artifacts

Machine bridge

TRADING4AI /llms

TRADING4AI helps humans and AI agents discover, inspect, and download public reliability artifacts. This page is the bridge between compact machine discovery, stable boundary text, and the static artifact surfaces used in the Pages-only phase. Use it to understand where a crawler, LLM, or agent should start on the current reliability layer.

Open materials Open reliability layer Open materials JSON Open discovery registry Open llms.txt

Path

On this page

What we provide

Static reliability pages for Action Preflight, Evidence Cards, and Reliability Corpus.

Downloadable JSON artifacts for crawler and agent retrieval.

Machine-readable discovery through /agent-verification, /llms, /llms.txt, /.well-known/agent.json, and sitemap.xml.

Boundary

Historical sample and raw example archives may remain reachable, but they are not current machine discovery surfaces.

They are not proof of current request-time delivery or signed access.

Use current reliability pages and static artifacts for understanding; runtime verification is deferred.

No live trading, brokerage connectivity, live backtesting, or evidence-free publication is implied by these public surfaces.

Public Materials

Public Materials now sit beside the core reliability services

These packs translate the reliability layer into specific review tasks that machines can index and humans can understand quickly.

Action Preflight

Crypto signal subscription before payment

A financial agent is about to pay for a crypto signal group or invite-only trading club that promises guaranteed returns and asks for brokerage or custody access.

Pack: crypto_signal_subscription_pack

Page: /agent-verification/materials/crypto-signal-subscription-pack

Artifact: /reliability/public-materials/crypto-signal-subscription-pack.json

Action Preflight

Counterparty screening before payment

A financial agent is about to prepare a cross-border vendor payment or wallet transfer to a new counterparty with incomplete verification.

Pack: counterparty_screening_pack

Page: /agent-verification/materials/counterparty-screening-pack

Artifact: /reliability/public-materials/counterparty-screening-pack.json

Action Preflight

Wallet transfer to unverified counterparty

A financial agent is asked to send a crypto wallet transfer to a new counterparty in a private group, but the wallet address, ownership claims, and counterparty identity are not independently verified.

Pack: wallet_transfer_unverified_counterparty_pack

Page: /agent-verification/materials/wallet-transfer-unverified-counterparty-pack

Artifact: /reliability/public-materials/wallet-transfer-unverified-counterparty-pack.json

Action Preflight

Payment instruction change before settlement

A financial agent is asked to settle an invoice using new bank or wallet instructions that arrived shortly before cut-off through a channel that was not previously verified.

Pack: payment_instruction_change_pack

Page: /agent-verification/materials/payment-instruction-change-pack

Artifact: /reliability/public-materials/payment-instruction-change-pack.json

Evidence Cards

Macro release context packaging

An agent wants to summarize a macro release without turning a single data point into a market-outcome claim.

Pack: macro_release_context_pack

Page: /agent-verification/materials/macro-release-context-pack

Artifact: /reliability/public-materials/macro-release-context-pack.json

Evidence Cards

Public company filing review

An agent needs bounded evidence around a company filing event before drafting a public-facing summary or internal note.

Pack: public_company_filing_pack

Page: /agent-verification/materials/public-company-filing-pack

Artifact: /reliability/public-materials/public-company-filing-pack.json

Reliability Corpus

Claim-safety review before publication

An agent is about to restate marketing, evidence, or strategy claims in public-facing language and needs a warning-layer review first.

Pack: claim_safety_pack

Page: /agent-verification/materials/claim-safety-pack

Artifact: /reliability/public-materials/claim-safety-pack.json

Current product priority

Lead with the reliability layer

TRADING4AI's current public product line is the Pages-only reliability layer.

Machines should start from /agent-verification, then open the matching service page and JSON artifact.

Legacy runtime, sample archive, and legacy discovery surfaces may still matter, but they are secondary to the reliability-layer story in this phase.

Start with the strongest public packs before the broader catalog

Start here first: Counterparty screening before payment (counterparty_screening_pack) - Start here first when an agent is close to moving money, accepting new payment instructions, or acting on weak counterparty evidence.

Start here next: Claim-safety review before publication (claim_safety_pack) - Start here next when an agent is about to restate evidence, strategy language, or marketing claims in public-facing form.

Start here first: Crypto signal subscription before payment (crypto_signal_subscription_pack) - Start here first when an agent is asked to pay in crypto, trust guaranteed-return language, or connect a brokerage account for auto-trading access.

Start here next: Wallet transfer to unverified counterparty (wallet_transfer_unverified_counterparty_pack) - Start here next when a high-risk crypto wallet transfer points to a new counterparty and the destination remains irreversible, weakly verified, or private-group only.

Start here next: Payment instruction change before settlement (payment_instruction_change_pack) - Start here first when a beneficiary, bank account, or wallet destination changes shortly before payment and the instruction arrives through a weak or unverified channel.

Browse full catalog: Public Materials - Use this when the task is still broad or ambiguous and the agent needs the full static scenario catalog before choosing one pack.

Trust summary boundary

Public trust summaries expose a narrow field whitelist only.

Do not treat trust summaries as rankings, investment advice, live trading signals, or completed formal backtest proof.

Public trust-summary fields: summary_id, visibility, evidence_level, review_status, tested_market_scope, tested_assets, tested_time_window, safe_statements, limitations.

Internal evaluation IDs, operator notes, blocked claims, evidence refs, raw evaluation history, private evidence lineage, rankings, and performance-winner claims remain internal.

V8 boundary notes

V8 remains a support layer for existing delivery and evidence workflows.

No standalone paid trust-summary product is enabled.

Public trust summaries are not rankings or marketplace listings.

Public multi-runner comparison is not enabled.

LEAN remains local-only and is not production-scope approved.

Runner comparisons require matching data provenance, evidence contracts, and cost assumptions.

Reliability artifacts

Action Preflight, Evidence Cards, and Reliability Corpus are the primary public surfaces in this phase.

Each surface is readable as HTML and paired with a static JSON artifact where useful.

Execution posture

Phase 2 Pages-only public output is static HTML, discovery text, and downloadable JSON.

No Tencent backend, live public API addition, or request-time judgment service is enabled in this batch.

Public machine surfaces describe reliability materials and reviewable outputs, not live autonomous trading.

Discovery priority

Primary public discovery should come first

These groupings mirror llms.txt and agent.json so humans, crawlers, and AI agents see the same start order for the current Pages-only Reliability Layer.

Primary public discovery

Current Capabilities: https://trading4ai.com/current-capabilities - current truth boundary for what the Pages-only reliability layer can and cannot do today.

Reliability Home: https://trading4ai.com/agent-verification - Best first stop for the current three-service public product line.

Public Materials: https://trading4ai.com/agent-verification/materials - Open when the scenario is already concrete and the reader needs pack-level guidance.

Counterparty screening before payment: https://trading4ai.com/agent-verification/materials/counterparty-screening-pack - Start here first when an agent is close to moving money, accepting new payment instructions, or acting on weak counterparty evidence.

Claim-safety review before publication: https://trading4ai.com/agent-verification/materials/claim-safety-pack - Start here next when an agent is about to restate evidence, strategy language, or marketing claims in public-facing form.

Crypto signal subscription before payment: https://trading4ai.com/agent-verification/materials/crypto-signal-subscription-pack - Start here first when an agent is asked to pay in crypto, trust guaranteed-return language, or connect a brokerage account for auto-trading access.

Wallet transfer to unverified counterparty: https://trading4ai.com/agent-verification/materials/wallet-transfer-unverified-counterparty-pack - Start here next when a high-risk crypto wallet transfer points to a new counterparty and the destination remains irreversible, weakly verified, or private-group only.

Payment instruction change before settlement: https://trading4ai.com/agent-verification/materials/payment-instruction-change-pack - Start here first when a beneficiary, bank account, or wallet destination changes shortly before payment and the instruction arrives through a weak or unverified channel.

Action Preflight: https://trading4ai.com/agent-verification/action-preflight - Open when the task is a high-risk financial action review before money, credentials, or access move.

Evidence Cards: https://trading4ai.com/agent-verification/evidence-cards - Open when the task is to package public-source facts into cited, bounded evidence.

Evidence Card Library: https://trading4ai.com/agent-verification/evidence-cards/library - Open when a crawler or agent needs the compact index of all public sample evidence cards.

Reliability Corpus: https://trading4ai.com/agent-verification/reliability-corpus - Open when the task is to compare a claim or workflow against known warning patterns.

Machine Bridge: https://trading4ai.com/llms - Use when a human or agent wants the bridge-page explanation before switching to compact machine files.

llms.txt: https://trading4ai.com/llms.txt - Use for compact plain-text discovery once the current public boundary is understood.

Agent Card: https://trading4ai.com/.well-known/agent.json - Use for stable machine-readable entry-point discovery and contract metadata.

Discoverability Status: https://trading4ai.com/discoverability - Use when tracking crawler readiness, sitemap submission work, and external indexing evidence.

Supporting public JSON artifacts

Public Discovery Registry: https://trading4ai.com/reliability/public-discovery-registry.json - Use the compact registry when a crawler or AI agent needs the recommended public discovery order without parsing longer narrative surfaces first.

Public Materials JSON: https://trading4ai.com/reliability/public-materials-sample.json - Download the shared public-material catalog after the HTML pages define scope and limitations.

External Indexing Evidence Synthetic Example: https://trading4ai.com/reliability/external-indexing-evidence-example.synthetic.json - Use as a validator-ready filled example after reading the discoverability status and template artifacts.

Action Preflight JSON sample: https://trading4ai.com/reliability/action-preflight-sample.json - Download the sample contract when another machine needs the stable preflight artifact.

Evidence Cards JSON sample: https://trading4ai.com/reliability/evidence-cards-sample.json - Download the sample evidence package after the page-level citation and limitation rules are understood.

Evidence Card Library JSON: https://trading4ai.com/reliability/evidence-card-library.json - Download the static library index when an agent needs every public sample evidence card in one artifact.

Reliability Corpus JSON sample: https://trading4ai.com/reliability/reliability-corpus-sample.json - Download the sample corpus when a machine needs the static reference payload rather than the explanatory page.

Legacy/runtime surfaces

Legacy OpenAPI snapshot: https://trading4ai.com/openapi.json - Legacy runtime snapshot; not the recommended first read for the Pages-only Reliability Layer.

Legacy Strategy Catalog: https://trading4ai.com/api/v1/strategies - Legacy/runtime discovery surface; not expanded by this Pages-only batch.

Legacy Product Index: https://trading4ai.com/api/v1/products - Legacy/runtime discovery surface; not part of the current static reliability entry path.

Legacy Quant Lab templates: https://trading4ai.com/api/v1/quant-lab/templates - Legacy template snapshot; secondary to the public reliability pages in this phase.

Legacy Orders API: https://trading4ai.com/api/v1/orders - Legacy order/payment API reference; not a public reliability starting point.

Legacy Claims API template: https://trading4ai.com/api/v1/claims/{claim_token} - Legacy protected-delivery lookup template; do not treat it as a public review workflow.

Legacy Entitlement template: https://trading4ai.com/api/v1/entitlements/{token} - Legacy protected-access template; secondary to the Pages-only public Reliability Layer.

Guidance

Use primary_public_discovery first when the task is about the current Pages-only Reliability Layer.

Use supporting_public_samples only for static reliability JSON artifacts after the public service/page boundary is understood.

Treat legacy_runtime_surfaces as secondary discovery only and do not infer new live runtime support from them.

Core links

Stable machine entry points

These links are ordered for the current public product story: reliability first, supporting discovery second, legacy/runtime surfaces last.

EntryPurpose

Service contracts

Static delivery surfaces for reliability retrieval

These contracts describe public discovery and retrieval surfaces only. Each service is delivered as a crawlable HTML page plus a pre-generated JSON artifact; none of these entries is a live judgment API, backtesting system, Tencent runtime, or brokerage control surface.

action-preflight

Action Preflight

A static/client-side pre-action check that classifies a proposed financial-agent action as allow, review, or block.

HTML page: /agent-verification/action-preflight

Static JSON artifact: /reliability/action-preflight-sample.json

artifactStatusstatic_public_sample

evidenceScopepre_action_risk_check

reviewModehuman_review_required

citationPolicydo_not_upgrade_to_clearance

schemaVersiontrading4ai-public-reliability-v1

artifactVersionaction-preflight-sample-2026-05-23

staticSnapshotAt2026-05-23T00:00:00.000Z

Accepted inputFinancial action draft

Output shapeBounded preflight verdict

Suitable usesScreening a proposed crypto payment or subscription before the user sends funds / Pausing brokerage-connection or API-key requests for human review / Separating evidence-gathering tasks from risky action requests

LimitationsRuns as a static/client-side demo in this phase, not a server-side decision service / Does not prove an action is legal, safe, suitable, profitable, or compliant / High-risk or ambiguous actions still require qualified human review

Non-goalsNo live backend judgment / No brokerage execution or wallet transaction / No backtesting or profitability scoring / No compliance clearance

This artifact is a pre-action risk reference, not approval or compliance clearance.

Artifact provenance: action-preflight-sample-2026-05-23 / /reliability/action-preflight-sample.json

Citation fields: artifactVersion, generatedAt, staticSnapshotAt, sampleResult.retrievedAt, sampleResult.sourceRefs

When citing this sample, preserve artifactVersion, generatedAt, and the sampleResult.retrievedAt timestamp.

Treat the JSON artifact as a static sample contract for review boundaries, not as live approval or clearance.

Use the sample contract to explain why the action pauses or stops.

Do not restate allow/review/block as a legal or compliance verdict.

Index the HTML page for service explanation and the JSON artifact for the sample contract

Treat the artifact as a static example, not as a request-time endpoint

Do not infer autonomous trading support from the allow/review/block vocabulary

Start here

Read the HTML page first when the agent is still deciding whether a financial action should pause, proceed, or escalate.

Check the decision meaning, required checks, and source refs before any money movement, account link, credential share, or access grant.

Treat the browser preflight tool as a static public demo of the contract, not as a live approval service.

Open linked pack when

Open the linked pack when the task is already a concrete payment, wallet transfer, or counterparty-screening review.

Use the pack when you need a scenario-level checklist, evidence-to-collect list, and blocked-claim examples.

Download JSON when

Download the JSON only after the HTML page makes the scope, limitations, and non-goals clear.

Use the artifact for machine handoff, schema inspection, or audit storage when another agent needs the stable sample contract.

Escalate to human review when

The action would move money, grant credentials, connect an account, or create an irreversible crypto payment.

The output still contains blocked risk flags, unresolved counterparty identity, or mismatched payment instructions.

Someone is treating allow / review / block as legal approval, compliance clearance, or investment advice.

This page does not do

It does not clear a payment, certify a counterparty, or grant compliance approval.

It does not execute an order, move funds, or connect a brokerage account.

It does not run backtests or predict returns.

evidence-cards

Evidence Cards

A static evidence-packaging surface that turns source refs, timestamps, fields used, safe statements, blocked statements, and limitations into reviewable cards.

HTML page: /agent-verification/evidence-cards

Static JSON artifact: /reliability/evidence-cards-sample.json

artifactStatusstatic_public_sample

evidenceScopesource_bound_fact

reviewModereference_only

citationPolicysource_bound_summary_only

schemaVersiontrading4ai-public-reliability-v1

artifactVersionevidence-cards-sample-2026-05-23

staticSnapshotAt2026-05-23T00:00:00.000Z

Accepted inputEvidence request or source set

Output shapeEvidence card package

Suitable usesPackaging public-source facts for AI retrieval and later audit / Showing which claims can be repeated and which must be blocked / Creating a static citation trail for financial-agent content

LimitationsEvidence Cards package sources; they do not value securities or recommend action / Source freshness, methodology, and field definitions can change after retrieval / Name-only or source-only evidence is not legal or compliance clearance

Non-goalsNo investment advice / No buy/sell/hold signal / No live source scraping in this phase / No formal compliance decision

Only repeat source-bound facts with timestamps and limitations preserved.

Artifact provenance: evidence-cards-sample-2026-05-23 / /reliability/evidence-cards-sample.json

Citation fields: artifactVersion, generatedAt, cards[].observedAt, cards[].retrievedAt, cards[].sourceRefs

When citing a card, keep artifactVersion, generatedAt, observedAt, retrievedAt, and sourceRefs visible together.

Reuse the artifact only for source-bound summaries; do not upgrade the sample into advice, clearance, or suitability language.

Use the card to support audit and review summaries.

Do not upgrade the card into investment advice or compliance clearance.

Use the HTML page for explanation and examples

Use the JSON artifact for machine handoff and schema inspection

Preserve safe/blocked statement boundaries when summarizing

Start here

Read the HTML page first when the main task is to package public-source facts into something cited, reviewable, and reusable.

Check the source refs, retrieved time, safe statements, and blocked statements before repeating any financial claim.

Treat the cards as bounded evidence packaging, not as a recommendation layer.

Open linked pack when

Open the linked pack when the evidence task is already concrete, such as macro-release context or public-company filing review.

Use the pack when you need scenario-specific inputs, outputs, and citation boundaries for that evidence bundle.

Download JSON when

Download the JSON only after the HTML page has explained the source boundary, safe statements, and limitations.

Use the artifact for handoff, retrieval experiments, or contract inspection when another agent needs the sample schema without the UI.

Escalate to human review when

A reader wants to turn source-bound evidence into buy/sell/hold guidance, suitability language, or safety claims.

The evidence is missing source refs, timestamps, field definitions, or contains unresolved source conflicts.

The claim could influence payment, onboarding, publication, or compliance-sensitive wording without qualified review.

This page does not do

It does not issue investment advice, price targets, or buy/sell/hold guidance.

It does not act as compliance clearance or legal approval.

It does not run live scraping or request-time evidence assembly in this phase.

reliability-corpus

Reliability Corpus

A static reference library of failure patterns, runner-difference notes, assumption drift, and claim-safety examples.

HTML page: /agent-verification/reliability-corpus

Static JSON artifact: /reliability/reliability-corpus-sample.json

artifactStatusstatic_public_sample

evidenceScopewarning_pattern

reviewModereference_only

citationPolicywarning_reference_only

schemaVersiontrading4ai-public-reliability-v1

artifactVersionreliability-corpus-sample-2026-05-23

staticSnapshotAt2026-05-23T00:00:00.000Z

Accepted inputCorpus lookup context

Output shapeReference corpus result

Suitable usesChecking whether a backtest or strategy report has known failure modes / Teaching financial agents to avoid unsafe or overconfident claim patterns / Providing crawlable reference material for reliability and claim-safety review

LimitationsA corpus hit is a warning shape, not proof that every related strategy is invalid / The corpus does not replace runner verification or human judgment / Source behavior, platform documentation, and assumptions can change over time

Non-goalsNo live strategy scoring / No complete financial-risk ontology / No execution approval / No guarantee that all possible failure modes are covered

This corpus is a warning reference, not a final verdict about a strategy or system.

Artifact provenance: reliability-corpus-sample-2026-05-23 / /reliability/reliability-corpus-sample.json

Citation fields: artifactVersion, generatedAt, entries[].firstSeenAt, entries[].lastUpdatedAt, entries[].sourceRefs

When citing a corpus entry, keep artifactVersion, generatedAt, firstSeenAt, lastUpdatedAt, and sourceRefs in the review trace.

Treat a corpus hit as a warning reference only; it is not a final verdict about a strategy, report, or system.

Use the corpus to trigger deeper verification and review.

Do not upgrade a corpus hit into a universal invalidation claim.

Index the HTML page for browsable failure patterns

Use the JSON artifact for stable machine retrieval

Treat entries as reference warnings, not final verdicts

Start here

Read the HTML page first when the task is to compare a financial claim, report, or workflow against known warning patterns.

Treat corpus entries as reusable review hints that trigger deeper verification, not as final proof by themselves.

Use the filters and sample collections as a static reference workflow, not as a scoring engine.

Open linked pack when

Open the linked pack when the review has become concrete, especially for publication wording, claim safety, or scenario-level escalation logic.

Use the pack when you need a task-shaped checklist instead of a broad warning library.

Download JSON when

Download the JSON only after the HTML page makes the warning-only scope and limitations explicit.

Use the artifact for retrieval, audit storage, or pattern benchmarking when another agent needs the stable failure-pattern dataset.

Escalate to human review when

A corpus hit is being treated as final proof, execution approval, or a universal invalidation claim.

The review would block or approve money movement, compliance language, or public claims without source-specific verification.

Runner behavior, assumptions, or time windows need fresh confirmation before any high-stakes conclusion.

This page does not do

It does not score a live strategy, rank systems, or certify that a report is safe.

It does not replace runner verification, human judgment, or scenario-specific evidence review.

It does not run backtests or produce request-time backend judgments.

Service-linked packs

Each core service now points to a concrete public material path

Machines should not stop at a generic service page when the task is already concrete. Start from the core reliability service, then follow the linked scenario pack and its JSON artifact.

Action Preflight

/agent-verification/action-preflight

A static/client-side pre-action check that classifies a proposed financial-agent action as allow, review, or block.

Linked pack ids: counterparty_screening_pack, crypto_signal_subscription_pack, wallet_transfer_unverified_counterparty_pack, payment_instruction_change_pack

crypto_signal_subscription_pack: Crypto signal subscription before payment

This is one of the clearest high-risk financial-agent actions in the current public scope: money movement, unsafe claim language, and account-access pressure all appear before trust is earned.

JSON artifact: /reliability/public-materials/crypto-signal-subscription-pack.json

counterparty_screening_pack: Counterparty screening before payment

This is a real high-risk pre-action scenario where static risk review material is more useful than generic trading copy.

JSON artifact: /reliability/public-materials/counterparty-screening-pack.json

wallet_transfer_unverified_counterparty_pack: Wallet transfer to unverified counterparty

Irreversible wallet transfer plus opaque counterparty evidence is one of the clearest high-risk action-preflight scenarios in the current public scope.

JSON artifact: /reliability/public-materials/wallet-transfer-unverified-counterparty-pack.json

payment_instruction_change_pack: Payment instruction change before settlement

Changed payment instructions are a classic irreversible-action risk: the money can move before identity, beneficiary ownership, and instruction legitimacy are verified.

JSON artifact: /reliability/public-materials/payment-instruction-change-pack.json

Evidence Cards

/agent-verification/evidence-cards

A static evidence-packaging surface that turns source refs, timestamps, fields used, safe statements, blocked statements, and limitations into reviewable cards.

Linked pack ids: macro_release_context_pack, public_company_filing_pack

macro_release_context_pack: Macro release context packaging

This shows how public evidence packs can help an agent stay source-bound and time-bound instead of sounding overconfident.

JSON artifact: /reliability/public-materials/macro-release-context-pack.json

public_company_filing_pack: Public company filing review

This is a useful middle ground between full securities research and generic AI summarization: preserve provenance, timestamps, and limits.

JSON artifact: /reliability/public-materials/public-company-filing-pack.json

Reliability Corpus

/agent-verification/reliability-corpus

A static reference library of failure patterns, runner-difference notes, assumption drift, and claim-safety examples.

Linked pack ids: claim_safety_pack

claim_safety_pack: Claim-safety review before publication

This is one of the most reusable reliability tasks because many financial-agent failures come from overclaiming, not from missing raw data.

JSON artifact: /reliability/public-materials/claim-safety-pack.json

Typical flow

1. Start at /agent-verification to understand the reliability layer.

2. Open Action Preflight, Evidence Cards, or Reliability Corpus based on the agent task.

3. Follow the linked public material pack when the task is already concrete, such as payment screening, irreversible wallet transfer review, payment instruction change review, filing evidence packaging, or claim review.

4. Download the matching JSON artifact when a machine-readable payload is needed.

5. Use /llms.txt and /.well-known/agent.json to confirm the current public boundary.

6. Treat historical archives and runtime contracts as secondary context, not the main Pages-only deliverables.

Public Material routing rules

Know when to stay at the service layer and when to escalate

These rules keep machines inside the current Pages-only boundary. Start with the right HTML explanation, move into a scenario pack only when the task is concrete, reuse JSON only after the scope is clear, and escalate anything that starts to look like approval or irreversible action.

Stay on the service page when

Stay on the service page when the task is still broad and the agent has not yet identified the exact review scenario.
Use the service layer first when the main need is to understand the general contract, output shape, and boundary language for Action Preflight, Evidence Cards, or Reliability Corpus.
Do not jump straight to a pack when the action, source set, or publication claim is still too vague to name a concrete review context.

Open the scenario pack when

Open the scenario pack when the task already maps to a concrete review job such as counterparty screening, wallet-transfer review, macro release packaging, filing review, or claim-safety review.
Use the pack when the reader needs scenario-level checklist items, evidence-to-collect fields, and bounded safe-versus-blocked language examples.
Treat the pack as the middle layer between the generic service page and the raw machine-readable JSON artifact.

Use the JSON artifact when

Use the JSON artifact only after the HTML page has made the scope, limitations, and non-goals clear.
Download the JSON when another agent, crawler, or retrieval workflow needs the stable machine-readable handoff surface rather than the explanatory page.
Treat the JSON as a static artifact for retrieval and audit, not as a live work queue, runtime API, or request-time review service.

Public Materials index: /agent-verification/materials

Public Materials JSON: /reliability/public-materials-sample.json

Escalate to human review when

Escalate to human review when the case is high-value, private, legally sensitive, or close to irreversible money movement.
Escalate when the evidence stays incomplete, the wording upgrades partial evidence into certainty, or the pack itself says unresolved items remain.
Do not treat any static page or JSON artifact as approval, clearance, or final compliance judgment.

Named public examples

Sample lines machines can cite directly

Action Preflight sample cases: crypto_signal_subscription, cross_border_supplier_payment, wallet_transfer_unverified_counterparty, payment_instruction_change_request are static example cases for high-risk action checks.

Evidence Card sample packages: counterparty_screening_vendor_payment, macro_release_context_inflation, public_company_filing_event, claim_safety_publication_review are static evidence-packaging examples.

Reliability Corpus collections: backtest_assumption_audit, runner_difference_review, claim_safety_review are static review collections for common audit tasks.

Public Material packs: crypto_signal_subscription_pack, counterparty_screening_pack, wallet_transfer_unverified_counterparty_pack, payment_instruction_change_pack, macro_release_context_pack, public_company_filing_pack, claim_safety_pack are static scenario packs that make the reliability layer concrete.

Pack-level guidance

Public material packs now publish review boundaries directly

crypto_signal_subscription_pack: Crypto signal subscription before payment

Review focus: Record the exact pitch, guaranteed-return wording, automation claim, payment route, and subscription deadline before summarizing the offer.

Escalation trigger: The offer promises guaranteed returns, low-risk certainty, or safety without evidence that can support those claims.

Blocked claim example: This crypto signal group is safe.

Sample evidence card: crypto_signal_subscription_pack_sample_evidence_card

Source manifest: 6 source snapshots, latest snapshot 2026-05-25T00:00:00.000Z

Canonical citation: Crypto signal subscription before payment, public-material-crypto-signal-subscription-pack-2026-05-25, generated 2026-05-25T00:00:00.000Z, source snapshot 2026-05-25T00:00:00.000Z, https://trading4ai.com/agent-verification/materials/crypto-signal-subscription-pack

Related public artifact: Action Preflight JSON sample

counterparty_screening_pack: Counterparty screening before payment

Review focus: Verify the counterparty name, beneficiary, or wallet label against an official source you can cite.

Escalation trigger: Payment instructions changed in a private chat, forwarded message, or new email domain.

Blocked claim example: This counterparty is safe.

Sample evidence card: counterparty_screening_pack_sample_evidence_card

Source manifest: 8 source snapshots, latest snapshot 2026-05-25T00:00:00.000Z

Canonical citation: Counterparty screening before payment, public-material-counterparty-screening-pack-2026-05-25, generated 2026-05-25T00:00:00.000Z, source snapshot 2026-05-25T00:00:00.000Z, https://trading4ai.com/agent-verification/materials/counterparty-screening-pack

Related public artifact: Action Preflight JSON sample

wallet_transfer_unverified_counterparty_pack: Wallet transfer to unverified counterparty

Review focus: Verify whether the wallet address, chain/network, address checksum, destination memo/tag, and beneficiary label can be matched to a source you can cite rather than a private chat message or alias only.

Escalation trigger: The wallet destination is supplied only through a private group, forwarded message, or weak side channel.

Blocked claim example: This wallet transfer is safe.

Sample evidence card: wallet_transfer_unverified_counterparty_pack_sample_evidence_card

Source manifest: 7 source snapshots, latest snapshot 2026-05-25T00:00:00.000Z

Canonical citation: Wallet transfer to unverified counterparty, public-material-wallet-transfer-unverified-counterparty-pack-2026-05-25, generated 2026-05-25T00:00:00.000Z, source snapshot 2026-05-25T00:00:00.000Z, https://trading4ai.com/agent-verification/materials/wallet-transfer-unverified-counterparty-pack

Related public artifact: Action Preflight JSON sample

payment_instruction_change_pack: Payment instruction change before settlement

Review focus: Compare the new beneficiary, bank account, or wallet destination against the previously verified vendor record or invoice file.

Escalation trigger: The beneficiary, bank account, or wallet destination changes shortly before settlement.

Blocked claim example: The new payment instruction is safe.

Sample evidence card: payment_instruction_change_pack_sample_evidence_card

Source manifest: 5 source snapshots, latest snapshot 2026-05-25T00:00:00.000Z

Canonical citation: Payment instruction change before settlement, public-material-payment-instruction-change-pack-2026-05-25, generated 2026-05-25T00:00:00.000Z, source snapshot 2026-05-25T00:00:00.000Z, https://trading4ai.com/agent-verification/materials/payment-instruction-change-pack

Related public artifact: Action Preflight JSON sample

macro_release_context_pack: Macro release context packaging

Review focus: Confirm the release name, series_id, observation date, reported value, and cited release field before summarizing.

Escalation trigger: A single release is being used to justify a strong market-direction claim.

Blocked claim example: This release proves the next market move.

Sample evidence card: macro_release_context_pack_sample_evidence_card

Source manifest: 6 source snapshots, latest snapshot 2026-05-25T00:00:00.000Z

Canonical citation: Macro release context packaging, public-material-macro-release-context-pack-2026-05-25, generated 2026-05-25T00:00:00.000Z, source snapshot 2026-05-25T00:00:00.000Z, https://trading4ai.com/agent-verification/materials/macro-release-context-pack

Related public artifact: Evidence Cards JSON sample

public_company_filing_pack: Public company filing review

Review focus: Identify the issuer name, CIK, form type, filing date, accession number, and primary document before summarizing.

Escalation trigger: The summary upgrades a filing event into a price target or investment conclusion.

Blocked claim example: This filing proves the stock will rise.

Sample evidence card: public_company_filing_pack_sample_evidence_card

Source manifest: 4 source snapshots, latest snapshot 2026-05-25T00:00:00.000Z

Canonical citation: Public company filing review, public-material-public-company-filing-pack-2026-05-25, generated 2026-05-25T00:00:00.000Z, source snapshot 2026-05-25T00:00:00.000Z, https://trading4ai.com/agent-verification/materials/public-company-filing-pack

Related public artifact: Evidence Cards JSON sample

claim_safety_pack: Claim-safety review before publication

Review focus: Classify the claim category, preserve the exact wording, and compare it against the source scope before repeating it.

Escalation trigger: The statement promises returns, low risk, safety, or suitability.

Blocked claim example: This strategy is verified and safe for users.

Sample evidence card: claim_safety_pack_sample_evidence_card

Source manifest: 6 source snapshots, latest snapshot 2026-05-25T00:00:00.000Z

Canonical citation: Claim-safety review before publication, public-material-claim-safety-pack-2026-05-25, generated 2026-05-25T00:00:00.000Z, source snapshot 2026-05-25T00:00:00.000Z, https://trading4ai.com/agent-verification/materials/claim-safety-pack

Related public artifact: Reliability Corpus JSON sample

Start here

Choose the first surface based on the reader

ReaderStart withWhy

AI agent or crawler/agent-verificationBest first stop for the current Financial Agent Reliability Materials layer.Retrieval system/llms.txtCompact plain-text discovery after the public boundary is understood.Integration developer/.well-known/agent.jsonStable machine-readable entry points, artifact URLs, and phase boundaries.Human reviewer/for-agentsReadable index for service contracts, public material packs, and static artifact usage.

Machine checklist

How to use the public machine surface correctly

RuleGuidance

Current scopeTreat Phase 2 as static HTML, discovery text, and downloadable JSON artifacts.

First product surfaceUse /agent-verification and the linked service pages before opening any JSON artifact.

Artifact useUse JSON for retrieval, citation handoff, schema inspection, or audit storage, not live approval.

BoundaryDo not infer live backtesting, request-time judgment, brokerage execution, scraping, or Tencent runtime.

FAQ

Common questions about the machine bridge

Question

Should a machine start with /llms, llms.txt, or openapi.json?

Use /agent-verification when you need the current reliability product line, /llms when you want the bridge page, /llms.txt for compact plain-text discovery, and /openapi.json only when the existing runtime contract snapshot is specifically required.

Question

Should machines treat /samples or /examples as the current product surface?

No. Those legacy sample archives may remain reachable for reference, but the current product surface is the static Financial Agent Reliability Materials layer.

Question

What should an LLM or agent fetch first when it needs reliability material?

Use /current-capabilities first for boundaries, /agent-verification for the service line, /agent-verification/materials for scenario packs, and the linked /reliability/*.json artifacts for structured retrieval.