Back to skill

Security audit

Dirextalk Deployer

Security checks for vulnerabilities and agentic risk

Overview

This skill is not clearly malicious, but it delegates AWS deployment and host skill changes to a mutable global npm package, so users should review it carefully before installing.

Install only if you trust the npm package publisher and are comfortable with a global, latest-version deployment tool changing your agent skill instructions and managing AWS resources. Prefer reviewing the exact npm package version and its SKILL.md, README.md, scripts, and package metadata first; use a least-privilege AWS identity and confirm any deploy, update, destroy, or skill-refresh action explicitly.

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

Error
Location
SKILL.md:19
Finding

Unpinned External npm Package Receives Global and Cloud-Capable Execution

Content
View full analysis

Vulnerability Details

File Location: SKILL.md, lines 19–28 and 40–49
Vulnerability Type: Unpinned third-party dependency and mutable supply-chain execution
Risk Level: High

Vulnerable Code:

markdown
This is the OpenClaw distribution of `dirextalk-deployer`. It is a thin
bootstrap skill: the executable deployment engine must always come from the
versioned npm package, not from a Git clone or this ClawHub bundle.

Before deployment, repair, verification, update, or teardown, install the
latest package:

```bash
npm install -g dirextalk-deployer@latest

Use the installed CLI to refresh the host-level skill when needed:

bash
dirextalk-deployer skill refresh --agent openclaw
text

```markdown
Read the npm package's `SKILL.md` and `README.md` before acting. They define the
required AWS billing confirmation, domain ownership confirmation, DNS handling,
security boundaries, and exact lifecycle commands.

The normal deployment entrypoint is:

```bash
DOMAIN=<domain> bash scripts/orchestrate.sh

Run it only from the installed deployer skill/runtime directory after the required confirmations. Use its status, verify, update, and destroy commands for existing services.

text

### Technical Analysis

The audited skill contains no local deployment implementation. Instead, it directs the user or agent to globally install `dirextalk-deployer@latest`, making a mutable external npm release the effective executable payload. The package version is not pinned, and the project provides no lockfile, integrity hash, provenance requirement, or locally reviewable copy of the executed scripts.

The installed package is also permitted to refresh the host-level OpenClaw skill and provide authoritative operational instructions through its own `SKILL.md` and `README.md`. Its orchestration script is expected to run in an AWS deployment context. Therefore, the behavior ultimatel
...[truncated 1955 chars]
Remediation
View remediation

Remediation Suggestions

  1. Replace dirextalk-deployer@latest with an exact, reviewed package version. Do not use a floating tag for security-sensitive deployment tooling.
  2. Record and verify the expected package integrity digest and npm provenance before installation.
  3. Include or vendor the exact SKILL.md, README.md, lifecycle scripts, package manifest, and lockfile that have been reviewed so the authorized behavior is visible in the audited artifact.
  4. Prefer a project-local, isolated installation over a global npm installation. Disable npm lifecycle scripts during installation unless they are explicitly required and reviewed.
  5. Require explicit review and confirmation before executing skill refresh, deployment, update, or destruction operations.
  6. Run the deployer with a dedicated least-privilege AWS identity. Separate deployment, verification, update, and destruction permissions where practical.
  7. Execute the package in an isolated environment with restricted filesystem, network, and credential access.
  8. Pin updates through a controlled release process that reviews source changes, dependency changes, signatures, and integrity metadata before changing the approved version.
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)

Ae1

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 44)May include surrounding context.

md
Read the npm package's `SKILL.md` and `README.md` before acting. They define the