XPR DeFi

Read-only DeFi queries for Metal X prices, swap rates, liquidity pools, and multisig proposal creation, approval, listing, and cancellation on XPR Network.

Do not provide your XPR private key to this skill until the author fixes inconsistencies and you fully understand the runtime behavior. Specific concerns: (1) The skill's manifest (skill.json) declares no required env vars, but the code requires XPR_PRIVATE_KEY, XPR_ACCOUNT, and XPR_RPC_ENDPOINT for any write operations. This means write tools will attempt to use your private key if you set those env vars — the skill should explicitly declare that requirement and mark write tools as optional. (2) The package dynamically imports '@proton/js' but dependencies aren't declared; ask the author to list dependencies or provide an install spec. (3) SKILL.md claims strong safety rules (confirmed:true, don't auto-execute proposals) — verify the runtime enforces them and that the agent platform will surface confirmation prompts. Recommended actions before installing: ask the publisher to update skill.json to list required env vars and dependencies, provide explicit documentation on when and how the private key is used, and make write tools opt-in; run the skill in an isolated test agent with no private key to validate read-only behavior; if you must enable writes, prefer using a time-locked or limited-signing key and keep the key out of broad environment scope (use ephemeral signing service or hardware wallet if possible).

Type: OpenClaw Skill Name: xpr-defi Version: 2.1.1 The skill is classified as suspicious due to its inherent high-risk capabilities, specifically the handling of a private key (`XPR_PRIVATE_KEY` from environment variables) for signing blockchain transactions and the `msig_propose` tool in `src/index.ts` which allows the agent to construct arbitrary blockchain actions for a multisig proposal. While these capabilities are necessary for a 'Complete DeFi toolkit' and are protected by strong safeguards like the `confirmed: true` parameter for all write operations and explicit prompt-injection prevention rules in `SKILL.md`, the potential for exploitation through a sophisticated prompt injection attack or a vulnerability in the agent's execution environment means it cannot be considered benign. There is no evidence of intentional malicious behavior such as data exfiltration or backdoor installation.