Stripe Projects Provisioning Skill
Provision third-party services, manage project resources, and retrieve credentials through the Stripe Projects CLI.
The skill constrains its tools and explicitly requires showing only environment-variable names and never printing secret values; it also uses an interactive prompt when a value is missing. However, it automatically runs --accept-tos --yes, installs plugins, and provisions third-party services, creating external effects without per-action confirmation, data-flow disclosure, permission scoping, rollback guidance, or dependency-security notes, so substantial points are deducted.
The workflow provides ordered commands, key checks, and recovery guidance for named error codes, including verification that the handoff skill was installed. Reliability still depends on the Stripe CLI, Projects plugin, and a later locally installed skill; there are no skill-specific tests or reproducible execution records, version and cross-platform handling are incomplete, and some failures shift diagnostics to the user, so the static ceiling and deductions apply.
Triggers cover common infrastructure requests such as databases, authentication, caching, hosting, credentials, monitoring, and service browsing, and vague requests are routed to catalog browsing. The trigger scope is nevertheless broad, non-fit boundaries and input/output contracts are underspecified, cancellation/read-only cases are unclear, installation guidance is mainly macOS-oriented, and Chinese-language or mainland-China reachability is not addressed, so points are deducted.
The document is layered across installation, catalog search, initialization, handoff, variable management, and error handling, with MIT licensing, update guidance, and repository synchronization signals. It lacks a named maintenance owner, changelog or explicit versioning, compatibility details, representative outputs, and comprehensive troubleshooting; its post-init capabilities also depend on a skill not included in the supplied evidence, so points are deducted.
The documented path plausibly covers catalog search, project initialization, service handoff, variable management, and a structured summary, making the core task potentially directly usable. Actual completion depends on external account state, CLI/plugin availability, a local follow-on skill, and a live catalog; there is no skill-specific verified output or comparison with alternatives, so the static score remains below the execution-based ceiling and is further reduced.
Commands, error codes, and repository CI/testing governance provide some auditable structure, and the README documents installation and synchronization sources. The supplied CI and tests do not cover this SKILL.md's key paths, and there is no third-party execution evidence or corroboration; the assessment therefore relies mainly on author-provided instructions and receives a low score.
- Obtain explicit user confirmation before running stripe projects init --accept-tos --yes, accepting terms, initializing a project, or creating external services.
- Because the workflow handles third-party credentials and environment variables, verify the CLI, plugin, follow-on skill sources, versions, permissions, and secret-storage boundaries.
- The core flow depends on Stripe Projects, external providers, and a live catalog; verify reachability in the target network, especially from mainland China.
- The supplied evidence does not demonstrate execution or testing of this SKILL.md's key paths; conclusions are documentation-level inferences.
What it does & when to use it
This Agent Skill from stripe/ai focuses on provisioning databases, authentication, hosting, caching, AI, monitoring, and other third-party services through the Stripe Projects CLI. It checks the Stripe CLI and Projects plugin, searches the service catalog, initializes a project, and hands the workflow to the locally installed stripe-projects-cli skill. It also covers project status, provisioned resources, environment variables, project variables, and provider connections. It fits developers who want centralized command-line provisioning, while actual availability depends on the Projects catalog, account eligibility, authentication, and terms acceptance.
Checks whether the Stripe CLI is installed and whether its version is at least 1.40.0; it provides installation or upgrade commands when needed. Installs the Projects plugin and searches the service catalog with stripe projects search or stripe projects catalog. Checks project status and runs preflight and initialization when a project is not initialized. Verifies the generated .claude/skills/stripe-projects-cli/SKILL.md, then hands off service addition, credential management, and project configuration to that skill. It also uses status, env, variables, and link commands to inspect resources, environment-variable names, project variables, and provider connections, while requiring secret values to stay out of responses.
- A developer needs Postgres, Redis, object storage, or a message queue and wants to confirm availability in the Projects catalog first.
- A team needs authentication, hosting, search, email sending, monitoring, or feature flags configured through Stripe Projects.
- A user needs an API key or credentials for a third-party service and wants the catalog checked before being told to register manually.
- A user is unsure which services are available and wants to browse the catalog before choosing a provider.
- A developer already using Stripe Projects needs to inspect project status, provisioned resources, or environment-variable names.
Pros & cons
- Covers the workflow from CLI checks and catalog search through project initialization.
- Documents handling for common blockers involving login, account eligibility, terms, OAuth linking, and plan dependencies.
- Treats the Projects CLI as the source of truth for project state and environment configuration, with explicit secret-handling guidance.
- Supports both targeted service searches and catalog browsing for vague requests.
- Requires Stripe CLI 1.40.0 or later and the Projects plugin.
- Provisioning may require Stripe login, account eligibility, terms acceptance, or provider OAuth linking.
- After initialization, the workflow depends on the locally generated `stripe-projects-cli` skill, whose detailed command reference is not included here.
- The source provides no test results and does not establish compatibility beyond the referenced agent harnesses.
How to install
The README documents installation for the repository's full skill collection, not a separate installation command for this skill. In a project, use codex plugin add stripe@openai-curated, claude plugin install stripe@claude-plugins-official, or npx skills add https://docs.stripe.com. This skill is located at skills/stripe-projects/SKILL.md; the source does not document a standalone command that installs only this skill. Manually installed skills do not auto-update; npx skills update -y updates them.
How to use
Trigger it for prompts such as “I need a database,” “set up auth,” “provision Redis,” “get an API key for a service,” “browse providers,” or any mention of projects.dev. First run which stripe && stripe --version and confirm Stripe CLI 1.40.0 or later; then run stripe plugin install projects. For a specific service, run stripe projects search <query> --json; for a vague request, run stripe projects catalog --json. Check stripe projects status --json. If the project is not initialized, run stripe projects init --preflight --json, then, when eligible, stripe projects init --accept-tos --yes. Verify .claude/skills/stripe-projects-cli/SKILL.md and hand off the remaining workflow.