Security audit
Security checks for vulnerabilities and agentic risk
This skill is a read-only post-trade review aid for Hyperliquid activity, with sensitive financial-record access that is disclosed and scoped to completed reviews.
Install only if you want an agent to review completed Hyperliquid activity. Provide only the account/history needed for the review, do not share private keys or signer secrets, and avoid running the optional unpinned `npx` handoff command unless you have verified and trust the target package/version.
Referenced artifact was not completely inspected
time, protection gaps, and result in the ticket's original risk unit. Use `node scripts/review.mjs --help` for simple round-trip arithmetic.
The documentation instructs use of npx skills add ... without pinning an exact package version or immutable artifact. That allows execution of whatever version is current at install time, creating a supply-chain risk where a compromised or malicious package update could run arbitrary code in the user's environment. In this skill context, the danger is somewhat reduced because installation is optional and the text tells restricted harnesses to present the command rather than execute it automatically, but it still normalizes an unsafe installation pattern.
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.
For HTTP-capable harnesses, use the official Info endpoint directly for reads. The equivalent metadata request is:
curl --fail-with-body https://api.hyperliquid.xyz/info \
-H 'Content-Type: application/json' \
--data '{"type":"meta"}'
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.
For HTTP-capable harnesses, use the official Info endpoint directly for reads. The equivalent metadata request is:
curl --fail-with-body https://api.hyperliquid.xyz/info \
-H 'Content-Type: application/json' \
--data '{"type":"meta"}'
The skill metadata promises review of Hyperliquid exchange records, fills, positions, funding, fees, and incidents, but the script only accepts user-supplied numeric CLI inputs and performs a local PnL calculation. This mismatch can mislead users or downstream agents into believing the output is based on authoritative exchange data, creating integrity and decision-making risk through false assurance rather than direct code execution abuse.
The default prompt says to use the skill to reconstruct 'this activity' from exchange records, but it does not clearly constrain when the skill should be invoked or explicitly exclude adjacent tasks like planning, live trading support, or broader market analysis. In an agentic environment, underspecified activation can cause the skill to be selected in the wrong context, potentially exposing sensitive trading records unnecessarily or producing misleading post-trade analysis outside its intended scope.
No suspicious patterns detected.