Back to skill

Security audit

Genexxis pay-per-call data APIs

Security checks for vulnerabilities and agentic risk

Overview

The skill is rated suspicious because scanners report repeated use of an unpinned `npx agentcash` command, which could run a changing package version; the scan was partial, so evidence is incomplete.

Review the skill's use of `npx agentcash` before installing, especially because the reported examples involve wallet or payment functions. The scan did not fully inspect the artifact, so the absence of other reported concerns should not be treated as evidence that it is safe.

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Tool MisuseTool Parameter Abuse, Chaining Abuse, Unsafe Defaults
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (7)

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
80% confidence
Finding

Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).

Content

Scanner excerpt · SKILL.md (reported line 116)May include surrounding context.

md
### Dockerfile, Compose & GitHub Actions linter (https://ciaudit.genexxis.com)

- `POST /compose` $0.01: Lint a docker-compose file: privileged, host network/pid/ipc, Docker socket mounts, SYS_ADMIN/ALL capabilities, unpinned images, database ports published on... Params: content*. [ciaudit_compose]
- `POST /dockerfile` $0.01: Lint a Dockerfile: base image pinning, root user, curl|sh, remote ADD, secrets in ENV/ARG, apt/pip/npm hygiene, chmod 777, EXPOSE 22, HEALTHCHECK, layer cache... Params: content*. [ciaudit_dockerfile]
- `POST /github-actions` $0.008: Lint a GitHub Actions workflow: actions not pinned to a commit SHA, missing or write-all permissions, pull_request_target with PR head checkout, script... Params: content*. [ciaudit_github_actions]

### Cloud & GPU price index (https://cloudprice.genexxis.com)

MCP server referenced without pinned version: 'npx agentcash'.

Medium
Category
MCP Rug Pull
Confidence
98% confidence
Finding

The accounts example uses agentcash@latest, exposing local execution to changes in the npm release selected at invocation time. The command concerns wallet accounts, making a compromised dependency particularly consequential; this is another occurrence of the same supply-chain weakness.

Content

No source excerpt is available for this finding.

MCP server referenced without pinned version: 'npx agentcash'.

Medium
Category
MCP Rug Pull
Confidence
98% confidence
Finding

The fetch command runs agentcash@latest through npx, allowing a mutable npm release to execute locally rather than using a reviewed, fixed version. Although this is a payment CLI rather than an MCP server, the underlying supply-chain concern is valid and heightened by the CLI's access to a funded wallet; payment caps do not constrain malicious package code.

Content

No source excerpt is available for this finding.

MCP server referenced without pinned version: 'npx agentcash'.

Medium
Category
MCP Rug Pull
Confidence
98% confidence
Finding

The price-check example also invokes agentcash@latest, so even a nominally free check can execute newly published package code locally. This repeats the same unpinned-dependency issue as the other commands on this line, not an independent MCP-server vulnerability.

Content

No source excerpt is available for this finding.

MCP server referenced without pinned version: 'npx agentcash'.

Medium
Category
MCP Rug Pull
Confidence
98% confidence
Finding

The sanctions-screening example executes agentcash@latest through npx, introducing mutable third-party code into the agent's local environment. User approval, an explicit payee, and a payment cap reduce unauthorized-payment risk during normal operation but do not prevent a compromised CLI from abusing local access.

Content

No source excerpt is available for this finding.

MCP server referenced without pinned version: 'npx agentcash'.

Medium
Category
MCP Rug Pull
Confidence
98% confidence
Finding

The spot-price example again invokes a mutable agentcash@latest release rather than a fixed dependency. A public-data query does not remove the local code-execution risk, especially when the same CLI manages payment credentials.

Content

No source excerpt is available for this finding.

MCP server referenced without pinned version: 'npx agentcash'.

Medium
Category
MCP Rug Pull
Confidence
98% confidence
Finding

The extraction example uses npx agentcash@latest, so running it can fetch and execute an unreviewed future release. The fixed public URL and capped payment are sensible safeguards, but neither mitigates malicious behavior inside the downloaded CLI.

Content

No source excerpt is available for this finding.