Back to skill

Security audit

CueCue Deep Research

Security checks for vulnerabilities and agentic risk

Overview

This financial research skill is coherent, but it can broadly route user finance questions to an external CLI that uses an API key, runs in the background, and writes persistent reports without enough user gating.

Install only if you trust CueCue and are comfortable sending financial research prompts, company details, and related context to its service. Prefer a local or isolated install over a global install, avoid putting confidential data in prompts unless approved, review where reports are written, and use a scoped API key that can be rotated.

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:25
Finding
Unpinned Third-Party Executable Dependency## Vulnerability Details **File Location**: `SKILL.md`, lines 25–41 **Vulnerability Type**: Unpinned npm executable dependency **Risk Level**: Medium ### Vulnerable Code ```yaml "requires": { "bins": ["node"], "env": ["CUECUE_API_KEY"] }, "install": [ { "id": "npm-global", "kind": "node", "label": "Install via npm (global)", "package": "@sensedealai/cuecue", "bins": ["cue"], }, { "id": "npm-local", "kind": "node", "label": "Install via npm (local)", "package": "@sensedealai/cuecue", "bins": ["cue"], }, ], ``` ### Technical Analysis The skill installs and executes the npm package `@sensedealai/cuecue` without specifying an exact version, integrity digest, lockfile, or immutable source reference. Consequently, installation can resolve to a package version that was not reviewed as part of this audit. The global installation option increases the potential effect because npm lifecycle scripts and the installed `cue` executable run with the installing user's privileges and may be available broadly in that user's environment. This repository does not include the dependency's source code, so its installation scripts, transitive dependencies, network behavior, credential handling, and runtime behavior cannot be verified here. This is a supply-chain weakness rather than evidence that the current package is malicious. ### Attack Path 1. An attacker compromises the npm publisher account, package release process, registry delivery path, or a transitive dependency. 2. The attacker publishes a malicious version that remains compatible with the unversioned package declaration. 3. A user or agent installs the skill dependency using either the local or global installation entry. 4. npm resolves and downloads the attacker-controlled release. 5. Malicious lifecycle code may execute during installation, or malicious behavior may run when the agent invokes `cue`. 6. The payload operates with the ...[truncated 870 chars]
Remediation
## Remediation Suggestions 1. Pin `@sensedealai/cuecue` to an exact, reviewed version rather than allowing resolution to the latest release. 2. Record and verify a cryptographic integrity digest for the approved package artifact. 3. Use a committed lockfile where the installation framework supports it, including pinned transitive dependencies. 4. Document and enforce the expected npm registry and canonical package source. 5. Prefer a local, isolated installation over a global installation to reduce command-path exposure. 6. Disable npm lifecycle scripts when they are unnecessary, or explicitly review every required lifecycle script before installation. 7. Audit the CLI source, its transitive dependencies, network destinations, and handling of `CUECUE_API_KEY` before deployment. 8. Run the CLI with least privilege in a restricted environment that exposes only required files and environment variables. 9. Add dependency monitoring and require manual security review before updating the approved package version.
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (4)

Vague Triggers

Medium
Confidence
96% confidence
Finding
The top-level description says the skill should be proactively invoked for essentially any data-backed financial analysis task, which is overly broad and can cause unnecessary routing of ordinary user conversations to an external research service. In this context, that increases the chance of sending sensitive prompts off-platform and performing file-writing side effects without clear user intent or consent.

Natural-Language Policy Violations

Medium
Confidence
89% confidence
Finding
The manifest description is entirely in Chinese and presents the skill as producing professional research output without indicating whether users can choose another language. This may violate language/locale policy when the skill is used in broader multilingual environments without explicit opt-in or documented regional limitation.

Missing User Warnings

Medium
Confidence
97% confidence
Finding
The skill documentation describes command usage but does not prominently warn that user queries are transmitted to an external service and that reports are written to local files. This omission prevents informed consent and can lead agents to disclose proprietary or personal financial information or create persistent local artifacts unexpectedly.

Vague Triggers

Medium
Confidence
96% confidence
Finding
The usage section explicitly instructs agents to proactively call the skill across a very wide range of scenarios, with no gating conditions, sensitivity checks, or requirement for user approval. Because this skill triggers external network access and file output, broad automatic invocation materially raises the risk of over-collection, privacy leakage, and unintended actions.

Static analysis

No suspicious patterns detected.