Jupiter Skill for OpenClaw
v1.0.2Execute Jupiter API operations on Solana - fetch quotes, sign transactions, execute swaps, prediction markets. Use when implementing token swaps, DCA, limit...
⭐ 2· 709·1 current·1 all-time
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Suspicious
medium confidencePurpose & Capability
The name/description (Jupiter API operations on Solana) match the included scripts (fetch-api, wallet-sign, execute-ultra, send-transaction). The capabilities requested by the scripts (JUP_API_KEY, access to a local Solana wallet file, optional SOLANA_RPC_URL) are appropriate for implementing swaps and signing/sending transactions. However, the registry metadata at the top of the submission lists no required env vars or config paths while SKILL.md and the code declare JUP_API_KEY and ~/.config/solana/id.json — this metadata omission is inconsistent and notable.
Instruction Scope
SKILL.md instructions and the scripts stick to the stated purpose: they call api.jup.ag endpoints, optionally a Solana RPC, read a local wallet JSON for signing, and serialize/send transactions. They do not attempt to read unrelated system files or exfiltrate secrets to unexpected endpoints. fetch-api accepts arbitrary Jupiter endpoints/bodies (expected for a general API helper).
Install Mechanism
There is no remote download/install step in the skill metadata; the repo includes package.json and a pnpm-lock.yaml and expects users to run pnpm install. Dependencies are standard npm packages (@solana/web3.js, commander, etc.), not downloads from untrusted URLs. This is a normal, moderate-risk Node dependency install (you will run code locally).
Credentials
The only secrets/access the scripts need are a Jupiter API key (JUP_API_KEY) and access to a local Solana wallet JSON (sensitive private key material). Those are proportionate to the stated purpose. The concern is the inconsistency: the registry metadata showed no required env vars or config paths while SKILL.md explicitly lists them. That omission could mislead users or automated installers about what secrets/files will be accessed.
Persistence & Privilege
The skill does not request always:true, does not modify other skills or system-wide configuration, and is user-invocable. It can be invoked autonomously by the agent (platform default), which increases blast radius if given access to wallet files — but autonomous invocation alone is not unusual and is not set as always:true.
What to consider before installing
This package appears to do what it says (talk to Jupiter API, sign and send Solana transactions), but take these precautions before installing:
- Verify the origin: the skill's source/homepage is unknown. Prefer code from a known repository or the official Jupiter examples.
- Metadata mismatch: SKILL.md and the scripts require JUP_API_KEY and access to a Solana wallet file (~/.config/solana/id.json), but the published registry metadata omitted these. Treat that discrepancy as a red flag — ask the publisher to correct it or verify why it was omitted.
- Wallet safety: never use a high-value wallet file with this tool. Use an isolated, low-balance or ephemeral wallet. Prefer hardware signing where possible; this code reads raw wallet JSON and will expose private key material to whatever runtime executes it.
- Run in a sandbox: installing dependencies (pnpm install) and running the scripts executes code locally. If you decide to test, do so in an isolated environment/container and inspect the code locally first.
- Inspect and verify: review the included scripts yourself (they are present and readable). Confirm endpoints are api.jup.ag and valid docs links; no other outbound endpoints are present. Check the pnpm-lock.yaml for unexpected packages if supply-chain trust is a concern.
- Autonomy risk: if you plan to allow the agent to invoke skills autonomously, be cautious about granting access to any wallet files or environment variables to the agent, because an autonomously-invoked skill could sign/send transactions without additional prompts.
If the publisher updates the registry metadata to include the required env vars and config paths and you verify the repo origin and contents, the skill would be coherently scoped for its purpose. If you cannot verify origin or the metadata remains inconsistent, avoid using real funds and prefer manual, audited usage only.Like a lobster shell, security has layers — review code before you run it.
latestvk975qmj28fjab6aree85rarkhn817889
License
MIT-0
Free to use, modify, and redistribute. No attribution required.
