Back to skill

Security audit

Blank Files Gateway

Security checks for vulnerabilities and agentic risk

Overview

This skill is a read-only helper for finding public blank test-file download URLs, with no artifact-backed evidence of hidden execution or data access.

Installing this skill should be acceptable for users who want public blank-file test downloads. Treat returned URLs as external downloads, and do not run the maintainer publishing commands unless you intentionally manage publishing and have verified the ClawHub CLI source/version yourself.

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
references/publish.md:8
Finding
Unpinned Global Installation of a Third-Party CLI## Vulnerability Details **File Location**: `references/publish.md`, lines 8-10 **Vulnerability Type**: Unpinned third-party dependency installed globally **Risk Level**: Medium **Vulnerable Code**: ```bash npm i -g clawhub clawhub login clawhub whoami ``` ### Technical Analysis The publishing instructions install the latest available version of the third-party `clawhub` npm package globally without pinning an audited version or verifying package integrity. An npm installation may execute package lifecycle scripts with the privileges of the invoking user. The global installation also changes system-wide user tooling rather than isolating the dependency to the project. Because the selected package version can change after this skill has been reviewed, compromise of the package, its dependency chain, or its publishing account could introduce attacker-controlled code. The subsequent authentication command may expose newly created ClawHub credentials or session data to a compromised CLI. These commands are part of maintainer publishing guidance rather than the skill's declared read-only file-discovery behavior, increasing the exposed supply-chain surface without being required during normal skill operation. ### Attack Path 1. An attacker compromises the `clawhub` npm package, one of its transitive dependencies, or an authorized package-publishing account. 2. The attacker publishes a malicious version or adds a malicious installation lifecycle script. 3. A maintainer follows `references/publish.md` and runs `npm i -g clawhub`, which resolves the mutable latest package version. 4. npm downloads the compromised package and may execute its lifecycle code with the maintainer's user privileges. 5. The maintainer then runs `clawhub login`; a compromised CLI can capture authentication material or misuse the authenticated session. 6. The malicious package can access resources available to that user and alter globally installed ...[truncated 486 chars]
Remediation
## Remediation Suggestions - Remove maintainer-only publishing instructions from the runtime skill package when they are not needed for normal operation. - Pin the CLI to a specific reviewed version, for example `clawhub@<exact-version>`, and update it only through a controlled review process. - Use the package manager lockfile and integrity metadata where applicable, and verify the package against official ClawHub documentation and publisher identity. - Prefer a project-local, isolated installation over `npm i -g` to reduce system-wide impact. - Inspect package contents and lifecycle scripts before installation. Where compatible with the trusted installation procedure, disable lifecycle scripts using `--ignore-scripts`. - Run publishing operations from a restricted environment with minimal filesystem access and short-lived, least-privilege credentials. - Separate installation and authentication, and verify the installed executable's provenance before running `clawhub login`.
Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • 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 (1)

Autonomous Decision Making

Medium
Category
Excessive Agency
Content
## Guardrails

- Treat this as a read-only gateway. Do not ask users to run shell scripts or installers.
- Do not fabricate file formats or URLs.
- Always verify availability via API before claiming a format exists.
- Use exact API route shapes (`/api/v1/...`), not deprecated routes.
Confidence
80% confidence
Finding
Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Static analysis

No suspicious patterns detected.