agentsclimarketplace

Revenue critical journey integrity review

Skill Raishin/vanguard-frontier-agentic/skills/cross-functional/revenue-critical-journey-integrity-review

Curated marketplace of AI skills, agents, and rules for cloud, zero-trust, and compliance-aware engineering - works with Claude Code, Codex, Cursor, Copilot, and more.

Install
npx -y skills add Raishin/vanguard-frontier-agentic --skill revenue-critical-journey-integrity-review

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 18 stars18 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Use this skill to review the cross-tier seams of revenue-critical journeys — checkout, payment submission, account creation, and login — for idempotency of money-moving and account-creating requests, server-side re-validation of client-enforced rules, webhook duplicate/out-of-order handling, retry-storm safeguards, and PCI DSS SAQ-scope judgment. Use when a request crosses client-to-server, system-to-processor, or webhook-back-into-system and a failure at that seam would double-charge, double-fulfill, bypass a required step, drop revenue, or misjudge PCI scope. Static review only; it does not execute payment flows and its PCI SAQ output is an advisory scoping opinion, never a compliance attestation.

SKILL.md

8.4 KB, ~1.7k tokens by cl100k_base, as published. Nobody here has run it

Revenue-Critical Journey Integrity Review

Purpose

Review the seams of revenue-critical journeys — the points where a request crosses from client to server, from your system to a payment processor, or from a webhook back into your system — so a journey that looks correct in any one tier does not break where the tiers meet. The dominant seam failures are non-idempotent money-moving requests, client-enforced rules the server never re-validates, webhook consumers that assume exactly-once/in-order delivery, unbounded retries that become retry storms, and PCI DSS SAQ-scope misjudgment.

When to use

Use this skill when the user asks to:

  • review whether a checkout, payment, subscription, coupon, or account-creation request is safe to retry (idempotency at money-moving seams),
  • confirm the server re-validates rules the client enforces (price, discount, quantity, eligibility, step-completion),
  • review a webhook consumer for duplicate-delivery and out-of-order handling,
  • review retry/backoff/circuit-breaker safety at a revenue-critical seam across web, mobile, or backend consumers,
  • get an advisory PCI DSS SAQ-scope opinion for the payment integration model actually in the code.

When not to use

Do not use this skill for:

  • tier-internal review that an owning specialist owns — DOM XSS/CSP and client injection (use the frontend security review), backend authorization-model design, mobile-platform specifics, or infrastructure hardening. This skill reviews the seam, not the interior; hand tier-internal findings to the owning agent.
  • issuing a PCI compliance attestation, signing an SAQ, or acting as an assessment of record. SAQ-scope output here is advisory only.
  • any live exercise of a payment system — executing flows, replaying webhooks, or sending requests to live/sandbox/staging processors. This skill is static review only.

Preconditions

  • The money-moving and account-creating request paths in scope, across whichever tiers exist.
  • The webhook consumer code and the event types it acts on.
  • The retry configuration (max attempts, backoff, jitter, timeout, circuit breaker) for the seams in scope.
  • The payment integration model (redirect, iframe/hosted fields, direct post/custom form) if a SAQ-scope opinion is requested.
  • The processor/SDK and version in scope, so idempotency and webhook guidance matches the real API surface.

Lean operating rules

  • Confirm retry/replay reachability before flagging an idempotency gap; a genuinely non-retryable internal call is not a finding.
  • Treat the server as the only enforcement boundary; a client-only check is UX, not enforcement.
  • Require webhook consumers to be both idempotent (dedupe by event id or business key) and order-tolerant.
  • Require bounded, backoff-with-jitter retries with a timeout at every revenue seam; add a circuit breaker or dead-letter path for backend/queue consumers.
  • Give a PCI SAQ-scope opinion only against the integration model present in the code, name the candidate SAQ, and label it advisory.
  • Never request, echo, store, or reproduce cardholder data, API keys, session tokens, or webhook signing secrets; redact-and-flag any that appear.
  • Label every claim repo evidence, context7-grounded, documentation-based, or inference.

Context7 documentation protocol

Processor idempotency semantics, webhook retry windows, event ordering, and signature verification are version-sensitive. The bundled official sources are the versioned ground truth for this skill: every processor-specific claim must trace to them and is labeled documentation-based, and the ledger records the version and last-verified date so a claim can be re-checked. This static-review skill's own tool grant is read-only (Read Grep Glob); when the invoking harness additionally provides Context7 or official-documentation tools, use them to confirm the current behavior against the bundled snapshot (resolve-library-id then query-docs, labeled context7-grounded) and to cover a processor the bundle does not. For a processor with no bundled or fetched coverage, say so and treat the claim as inference — never rely on memorized API details.

Workflow

Follow the step-by-step review and output contract in workflow and output. At a high level: (1) map the seams in scope; (2) for each money-moving/account-creating request, check idempotency against reachable retry/replay; (3) check server re-validation of every client-enforced rule; (4) check webhook consumers for idempotency + order-tolerance; (5) check retry safety; (6) if requested, form the advisory SAQ-scope opinion; (7) emit findings with evidence tiers and tier-internal handoffs.

Decision gates

  • Block only on a seam failure with a demonstrated reachable retry/replay/bypass path.
  • Every processor-specific claim is Context7-grounded or documentation-based, never memory.
  • Every SAQ-scope statement is advisory and tied to the integration model in the code.
  • Every tier-internal finding is handed off, not adjudicated.

Evidence classification

Label each finding repo evidence (seen in the code), context7-grounded (current provider docs via Context7), documentation-based (official docs), or inference. Documentation never proves a specific deployment's live behavior — say so.

Security and privacy constraints

Static review only. Never transmit, request, store, or reproduce cardholder data (PAN/CVV), API keys, session tokens, or webhook signing secrets; treat any such string as a redact-and-flag finding. Never execute payment flows or contact any live/sandbox/staging payment system. PCI SAQ-scope output is an advisory opinion, never an attestation.

Escalation conditions

Escalate to incident response on any evidence of a live failure (duplicate charges in logs, replayed webhooks, retry amplification). Escalate SAQ-scope opinions to the merchant's compliance owner or a QSA as advisory input.

References

Load these only when needed:

Response minimum

Return, at minimum:

  • the seam(s) in scope and, per finding, the failure class and evidence tier;
  • the cross-tier failure narrative (how a retry, replay, or bypass reaches a wrong outcome);
  • concrete remediation and an exact verification step;
  • the advisory SAQ-scope opinion when requested, labeled advisory;
  • tier-internal handoffs and any incident-response escalation.

Anti-goals

  • Do not expand into tier-internal review; own the seam, hand off the interior.
  • Do not present a SAQ-scope opinion as a compliance determination.
  • Do not exercise any live payment system or reproduce any secret or PAN.
  • Do not assert processor behavior from memory.

What ships with it: 7 files

47.7 KB alongside SKILL.md

Keep looking

Skills are one crate of 327,069. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.