MnemoPay

Give any AI agent persistent memory and a wallet. Remember facts across sessions, recall with semantic search, charge for work delivered, track reputation. 1...

Do not install blindly. Key things to verify before using this skill: - Source provenance: confirm the npm package and GitHub repo are legitimate, review the package code and maintainer identity. - Installation risk: the SKILL.md tells the agent to run `npx @mnemopay/sdk` — that will fetch and execute remote code. Only allow this in a sandboxed environment after code review. - Credentials and infrastructure: clarify what DB, Redis, and payment credentials are required and ensure they are provided via secure secret storage (not hard-coded). Ask how payment escrow is implemented and what third-party gateways are used. - Autonomy and billing safeguards: require explicit human approval before any charge is created or settled; restrict autonomous invocation for payment-related tools; add limits and audit hooks. - Data governance: ask how memories and logs are stored, encrypted, retained, and deleted (privacy/PDPA/GDPR concerns). If you cannot validate the package source and the payment/back-end design, treat this skill as unsafe for production and run only in an isolated test environment with strict limits on network and financial actions.

Type: OpenClaw Skill Name: mnemopay Version: 1.1.0 The skill introduces high-risk capabilities by providing tools for financial transactions (`charge`, `settle`) and persistent memory management. It executes remote code via `npx -y @mnemopay/sdk` and includes specific instructions in `SKILL.md` for the agent to initiate charges and manage a wallet. While these features are consistent with the stated purpose, the combination of automated financial workflows and the use of an atypical GitHub repository source (`t49qnsx7qt-kpanks/mnemopay-sdk`) warrants a suspicious classification due to the potential for financial abuse or supply chain risk.