Back to skill

Security audit

Research Brief Generator

Security checks for vulnerabilities and agentic risk

Overview

This is a plain Markdown research-brief skill with broad but disclosed topic coverage and no hidden code, credential use, persistence, or data access behavior.

Install only if you are comfortable with the documented ClawHub npx installer flow; for stricter supply-chain control, use a pinned and trusted installer version. Expect the skill to activate for intentionally requested research briefs on many kinds of topics, including sensitive subjects where it may produce content warnings.

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
README.md:37
Finding
Unpinned Executable npm Dependency in Installation Command## Vulnerability Details **File Location**: `README.md`, line 37 **Vulnerability Type**: Supply-chain risk from a mutable executable dependency **Risk Level**: Medium ### Vulnerable Code ```bash npx clawhub@latest install tetsuakira-vk/research-brief ``` ### Technical Analysis The documented installation command uses `npx` to download and execute the mutable `latest` release of the `clawhub` npm package. No exact package version, package integrity hash, lockfile, or other verification mechanism is specified. Because the `latest` distribution tag can be reassigned after this project has been reviewed, the code ultimately executed by users may differ from the version that was available during the audit. If the upstream package, maintainer account, publishing credentials, or package registry is compromised, a malicious release could be served through this command. This is classified as an insecure dependency rather than a confirmed malicious payload: the reviewed project contains no malicious script, and exploitation depends on a future or existing compromise of the external dependency or its distribution channel. ### Attack Path 1. An attacker compromises the `clawhub` npm package, its maintainer account, publishing credentials, or the relevant registry distribution channel. 2. The attacker publishes a malicious package version and assigns it to the `latest` distribution tag. 3. A user follows the installation instructions and runs the documented `npx clawhub@latest ...` command. 4. `npx` retrieves and executes the attacker-controlled package code. 5. The malicious package operates with the privileges and environment access of the user who launched the command. ### Impact Assessment Successful exploitation could permit arbitrary code execution under the installing user's account. Depending on that user's privileges and environment, the payload could access user-readable files, environment variables, tokens, development credentials, source repositories ...[truncated 338 chars]
Remediation
## Remediation Suggestions 1. Replace `@latest` with an exact, reviewed version: ```bash npx clawhub@X.Y.Z install tetsuakira-vk/research-brief ``` 2. Avoid semver ranges and mutable distribution tags in security-sensitive installation instructions. 3. Verify the selected package version and provenance before documenting it. 4. Where supported, validate the package against an expected registry integrity hash or trusted signature. 5. Use a lockfile or an equivalent dependency manifest when installation is performed through a managed project. 6. Document the expected npm registry and trusted package owner to reduce dependency-confusion and registry-substitution risks. 7. Establish an update process in which new installer versions are reviewed before the pinned version is changed.
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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
Findings (3)

Rp1

Medium
Category
MCP Rug Pull
Confidence
93% confidence
Finding
The installation command uses `npx clawhub@latest`, which pulls and executes the latest published package version at install time rather than a pinned, reviewed release. That creates a supply-chain risk: if the upstream package is compromised or a breaking/malicious version is published, users may execute untrusted code simply by following the README.

Vague Triggers

Medium
Confidence
89% confidence
Finding
The trigger phrase `Use research-brief on: [topic, case, person, or event]` is broad and resembles ordinary natural-language conversation, making accidental or unintended invocation more likely. In an agent environment, loose activation can cause the skill to run when a user is merely discussing a topic, which may override expected behavior, consume resources, or cause the agent to produce research output in contexts where it was not explicitly requested.

Vague Triggers

Medium
Confidence
91% confidence
Finding
Line L017 says to 'Accept any topic' and then lists very broad categories, while the front-matter description also frames the skill as usable for 'any topic, person, case, or event.' For a markdown skill file, this is an ambiguous trigger scope with little constraint or exclusion guidance, increasing the chance of unintended invocation for general requests.

Static analysis

No suspicious patterns detected.