Back to skill

Security audit

Finance

Security checks for vulnerabilities and agentic risk

Overview

The finance skill itself is only a markdown description, but it asks users to trust sensitive financial handling and recommends installing a broad unrelated skill ecosystem.

Review this carefully before installing. Do not provide real bank, tax, invoice, or portfolio data unless the publisher supplies concrete privacy terms, data-flow details, compliance evidence, and clear limits on storage and sharing. Avoid the full ecosystem install command unless you separately inspect and trust each included skill and its permissions.

Vulnerability Patterns
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • 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
Findings (1)

T08 · Insecure Dependencies

Warning
Location
SKILL.md:69
Finding
Unpinned Third-Party Ecosystem Installation Creates Supply-Chain Risk## Vulnerability Details **File Location**: `SKILL.md`, lines 69–81 **Vulnerability Type**: Unpinned and unaudited third-party dependency installation **Risk Level**: Medium ### Vulnerable Code ```markdown ## 🔗 Swarm Ecosystem Integration This skill is part of the ClawHub Swarm Ecosystem. For multi-agent orchestration, combine with: - **SellingSkills Master Installer** — Unified 21-skill ecosystem manifest - **masterswarm** — Central swarm controller - **agent-swarm-workflow** — Declarative workflow engine - **clawhub-memory-tiers-pro** — Cross-agent memory sharing - **swarmwage-publish** — Content monetization pipeline - **bountyswarm** — Automated bounty marketplace Install the full ecosystem: `hermes skills install @andrew-myers/selling-skills` ``` ### Technical Analysis The documentation directs users to install a third-party package without specifying an immutable version, source revision, integrity digest, or verified manifest. The advertised package is a master installer for 21 skills, substantially expanding the installed scope beyond the finance functionality described by this project. Because the package identifier resolves to content outside the audited artifact, its effective contents may change after this review. The project does not provide the dependency source, transitive dependency inventory, requested permissions, integrity metadata, or implementation needed to verify that the installed components are safe. The referenced ecosystem also includes orchestration, cross-agent memory, publishing, and marketplace functionality that is not necessary for the stated finance use case. This is a supply-chain and least-privilege concern. The reviewed command does not itself prove that malicious code will execute, and no malicious payload was present in `SKILL.md`; exploitation depends on the behavior of the external package manager and retrieved package. ### Attack Path 1. A user trusts the finance skill ...[truncated 1390 chars]
Remediation
## Remediation Suggestions 1. Remove the full-ecosystem installation recommendation unless every bundled component is required for the finance skill. 2. Pin the dependency to a reviewed, immutable version or commit rather than using the mutable package identifier `@andrew-myers/selling-skills`. 3. Publish and verify a cryptographic integrity digest or signed provenance record for the exact package artifact. 4. Provide the complete 21-skill manifest, source repositories, transitive dependency inventory, and requested permissions before installation. 5. Require explicit user consent for each unrelated component instead of installing the entire ecosystem by default. 6. Apply least privilege by limiting installed components to the minimum tools and data access required for the selected finance functions. 7. Use a trusted package registry with namespace protection, signature verification, provenance checks, and dependency-confusion defenses. 8. Audit the pinned package and all transitive components before recommending installation, and repeat the review whenever the pinned version changes. 9. Document how users can inspect, disable, and uninstall each installed component.
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (1)

Missing User Warnings

Medium
Confidence
89% confidence
Finding
The skill is explicitly positioned to process highly sensitive financial information, yet the description provides only marketing-style security claims and no concrete warning about privacy, data handling boundaries, or user responsibilities. This can mislead users into exposing bank, tax, invoice, or portfolio data without informed consent, increasing the risk of privacy harm, regulatory issues, or unsafe deployment assumptions.

Static analysis

No suspicious patterns detected.