← 返回 Skills 市场
321
总下载
0
收藏
1
当前安装
3
版本数
在 OpenClaw 中安装
/install raven-transfer
功能描述
Wallet-aware Raven Atlas transfer operations for NGN payouts. Use when an agent must check wallet balance, resolve Nigerian bank accounts, enforce explicit c...
安全使用建议
This package appears to implement legitimate Raven NGN payout operations and correctly requires a Raven API key and a local state file for idempotency. However: (1) the registry metadata shown with the published skill incorrectly lists no required environment variables and no primary credential — confirm that your agent runtime will supply RAVEN_API_KEY_FILE or RAVEN_API_KEY before installing; (2) SKILL.md and agents/openai.yaml ask that implicit model invocation be disabled to prevent accidental autonomous transfers — make sure the skill is not allowed to run implicitly/autonomously in your agent configuration; (3) review and exclude scripts/.state/transfer-state.json from backups and source control (the package claims it stores only non‑PII idempotency fields, but verify you agree); (4) run the included validate-skill-package.mjs and unit tests in a safe environment (live transfer tests are intentionally gated) before enabling; (5) treat this as high-impact software — it can move money, so if you cannot verify the registry metadata and agent policies match the package docs, do not install or provide production API keys. If the publisher can correct the registry metadata (declare the required env and primary credential and set disable-model-invocation), confidence in installing this skill would increase.
功能分析
Type: OpenClaw Skill
Name: raven-transfer
Version: 1.1.0
The raven-transfer skill bundle is a well-architected tool for managing NGN payouts via the Raven Atlas API. It includes robust safety features such as mandatory confirmation tokens to prevent accidental execution, idempotency checks via local state to prevent double-spending, and strict file permission enforcement (chmod 600) for API key storage. The implementation in scripts/raven-transfer.mjs demonstrates high security awareness by explicitly stripping PII (account numbers/names) before persisting state and implementing comprehensive log redaction for sensitive tokens and headers.
能力评估
Purpose & Capability
The code, README, SKILL.md, and agent metadata all implement wallet-checked NGN transfers via Raven Atlas and require a Raven API key; that capability matches the skill name and description. However, the registry-level metadata claims no required environment variables or primary credential while the package clearly requires RAVEN_API_KEY_FILE or RAVEN_API_KEY. This inconsistency is unexpected and should be resolved.
Instruction Scope
SKILL.md and the scripts only instruct operations needed for payouts (balance, lookup, preview, confirm, execute, status). The runtime instructions reference only the Raven API and a local state file (.state/transfer-state.json). They explicitly forbid auto-retry for transfers and require explicit confirmation tokens, which is in-scope for the described purpose.
Install Mechanism
There is no remote install/download step; this is an instruction + local script package. All code is included in the bundle and there are no external URLs or archive extraction steps. Risk from install mechanism is low.
Credentials
The package legitimately requires a Raven API credential (file or env var) and several optional test/run-time overrides. Those are proportionate to payout functionality. The concern is that the registry metadata provided alongside the skill claims no required env vars/primary credential, conflicting with agents/openai.yaml and SKILL.md; that mismatch could lead to accidental missing protections or incorrect installation by users/automation.
Persistence & Privilege
The script persists a minimal idempotency state to scripts/.state/transfer-state.json with owner-only permissions by default and provides RAVEN_DISABLE_LOCAL_STATE to disable persistence. This is consistent with its purpose; ensure the .state directory is excluded from backups and VCS. The skill does not request always:true or other elevated platform privileges.
如何使用
- 确保已安装 OpenClaw（本地或 Docker 部署）
- 在对话框中输入安装命令：
/install raven-transfer - 安装完成后，直接呼叫该 Skill 的名称或使用
/raven-transfer触发 - 根据 Skill 的参数说明提供必要输入，即可获得结构化输出
版本历史
v1.1.0
- Added CLI implementation script (`scripts/raven-transfer.mjs`) and supporting scripts for validation and testing.
- Introduced `agents/openai.yaml` for runtime environment metadata.
- Expanded test coverage with live contract and unit tests.
- Updated documentation to reflect new package structure (`README.md`, process, and artifacts).
- No changes to the core transfer process or environment requirements.
v1.0.1
- Improved documentation for failed transfer handling: added guidance on auto-reversals and retry checks.
- Updated environment requirements: now prefers RAVEN_API_KEY_FILE and documents additional runtime safety, timeout, and endpoint options.
- Added section on invocation policy, clarifying explicit invocation and disabling implicit model use.
- Expanded configuration details and usage notes throughout reference files.
- Removed unused scripts, agent config, and test files for a leaner skill footprint.
v1.0.0
raven-transfer 1.0.0 – Initial release
- Introduces wallet-aware transfer operations for NGN payouts via Raven Atlas.
- Ensures checks for wallet balance and explicit transfer confirmation tokens before execution.
- Supports Nigerian bank account resolution and merchant settlement transfers.
- Provides normalized results and strict idempotency, with no auto-retries.
- Requires `RAVEN_API_KEY` in runtime environment for operation.
元数据
常见问题
Raven Transfer 是什么？
Wallet-aware Raven Atlas transfer operations for NGN payouts. Use when an agent must check wallet balance, resolve Nigerian bank accounts, enforce explicit c... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 321 次。
如何安装 Raven Transfer？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install raven-transfer」即可一键安装，无需额外配置。
Raven Transfer 是免费的吗？
是的，Raven Transfer 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
Raven Transfer 支持哪些平台？
Raven Transfer 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（cross-platform）。
谁开发了 Raven Transfer？
由 Fmac（@fmacmak）开发并维护，当前版本 v1.1.0。
推荐 Skills