Stripe AI is a collection of SDKs, tools, an MCP server, and agent skills for building AI products and businesses with Stripe's billing infrastructure. Developers use it to connect Stripe payments and metering to language models and agent frameworks, while the catalogue entries provide Stripe-specific skills, commands, plugins, and agent integrations.
Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add stripe/ai --skill connect-required-verification-informationgit clone --depth 1 https://github.com/stripe/aiWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/stripe/ai/connect-required-verification-information)<a href="https://agentmods.dev/skills/stripe/ai/connect-required-verification-information"><img src="https://agentmods.dev/badge/skills/stripe/ai/connect-required-verification-information/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/stripe/ai/connect-required-verification-information"><img src="https://agentmods.dev/badge/skills/stripe/ai/connect-required-verification-information.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- Socket pass
- Snyk warn
- NVIDIA SkillSpector warn
SkillSpector: 1 finding, up to high
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- high YARA Match · line 34 YARA rule matched a hack tool or exploit indicator (offensive tools, reconnaissance, privilege escalation, or exploit frameworks).Fix: Remove offensive tool references and exploit code. Legitimate agent skills should not contain penetration testing tools, exploit frameworks, or reconnaissance utilities.
What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00080 | $0.06327 |
| Opus 5 | $0.00040 | $0.03164 |
| Sonnet 5 | $0.00016 | $0.01265 |
| Haiku 4.5 | $0.00008 | $0.00633 |
Grade A, and why
connect-required-verification-information scanned grade A with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 10d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
### curl examples How it starts
The opening of the file, as written. The whole thing — 413 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Instructions
The human-accessible version of this documentation allows the user to select connected account fields and regions using a form, and then makes API requests to fetch and display the requirements a connected account with the selected configuration and region must provide. Follow these instructions to fetch the same information.
Interaction contract
Terminology used in this document:
field: a setup input such asplatformCountry,accountCountry, orcapabilitiesoption: a presented selectable option for a fieldvalue: the option the user selects, or the free-response value the user provides for a field
Every time you ask the user to provide a value for a field:
- use a multiple-choice question; never stop at a plain free-form prompt or wait for raw chat input
- if you need free user input, instruct the user to use the question’s free-response field
- for long option lists, explicitly say that any value from the full validated list is still accepted through the free-response field
- if the user already provided a valid answer in an earlier message, use that instead of asking again
Hard rules
You must follow these rules:
- Ask for a field only after all of its prerequisite fields are satisfied.
- Collect setup fields progressively as the flow advances.
- Ask for one field at a time, or one group of fields only when they are dependency-free at that point in the flow.
- For example, ask for
platformCountryandaccountCountryseparately: the platform country determines which account countries are valid, so asking both together can produce invalid combinations. But you may ask fordashboardType,tosType, andlegalEntityTypetogether in one group because their valid options are already known from the same response.
- For example, ask for
- If there is ever a conflict between the user’s request and the validated setup, inform the user of the conflict and ask them to revise their setup choices using the Interaction contract. Keep the validated setup aligned with what the user requested without silently dropping the conflict.
- Follow the Interaction contract for every user question.
- When the number of available options exceeds four, always print the full validated reference list before asking the multiple-choice question so the user can see the full option space.
- When printing countries, always print the full country name followed by its code in parentheses, for example,
Germany (DE). - In the multiple-choice question, include a small set of suggested options so the user can move forward with immediate clarity. The reference list above remains the authoritative full set.
- Leave the descriptions for the country suggested options blank.
- When printing countries, always print the full country name followed by its code in parentheses, for example,
- For any field with four or fewer valid options, show every valid option directly in the multiple-choice question. Do not print a separate reference list first.
- Every list of selectable options shown to the user must be pre-validated against all currently known constraints before you display it.
- Never display an option as selectable if you already know it will be removed, rejected, or auto-adjusted later in the flow.
- Present options that stay valid through the current flow.
- Ask about
capabilitiesafterplatformCountry,accountCountry, and the downstream validity constraints for that setup are resolved. - Only ask about
orrProgramwhen it is present in the publicprogramsreturned for the validated setup. - If the
businessStructuremap for the chosenlegalEntityTypeis empty or contains exactly one keynil, skipbusinessStructure. Otherwise, ask forbusinessStructureand always allow anoneoption or leave unselected as a suggested option in the multiple-choice question. - If the user decides to change an earlier choice like
platformCountry, you must invalidate and re-check all downstream fields before continuing. - Keep the dependency chain implicit. Share the information the user needs to make progress and keep the experience simple.
- Use external-facing language when talking to the user. See below to translate the internal API terminology.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 10d ago First seen · 413 lines · 80 tokens per session scan A a9c30050a332
connect-required-verification-information is a skill published in the GitHub repository stripe/ai (1,795 stars, last pushed today), licensed MIT. It adds 80 tokens to every session and 6,327 once invoked, about $0.0004 per session on Opus 5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
ax-python-llm
Use when writing Python code with axllm for using the generated Ax package, factory functions, package docs, examples, and API reference.
ax-python-signature
Use when writing Python code with axllm for string signatures, field descriptors, JSON schema output, validation, and typed tool argument shapes.
deepseek-harness
Use when building AI agent applications with a plugin-based architecture — Web UI, CLI, Python SDK, Cordis plugin system, multi-model orchestration. DeepSeek Harness (dsh): open-source agent harness by DeepSeek AI where everything is a plugin, powered by Cordis for spatiotemporal composability.
ax-python-gepa
Use when writing Python code with axllm for GEPA, Pareto tradeoffs, reflection clients, metric budgets, optimizer state, and artifacts.
model-context
MCP (Model Context Protocol) - Build AI-native servers with tools, resources, and prompts. TypeScript/Python SDKs for Claude Desktop integration.
ax-python-gen
Use when writing Python code with axllm for AxGen programs, forward calls, indexed multi-sampling, result pickers, streaming, tools, assertions, traces, usage, and output parsing.