ClawHub
Back to skill

Security audit

productOptimizer

Security checks for vulnerabilities and agentic risk

Overview

The skill is a real BigCommerce catalog updater, but it can autonomously rewrite live product content across a store with weak confirmation, validation, and credential-handling safeguards.

Install only if you are prepared for it to modify live BigCommerce product content in bulk. Use a test store first, export or back up existing catalog content, use a narrowly scoped token, avoid putting tokens on the command line, review every generated update file before pushing it, and consider replacing the setup with a pinned virtual environment. This is not evidence of malicious intent, but the current safeguards are weak for production catalog writes.

A.I.G

Vulnerability Patterns
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • 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
Findings (3)

T09 · Insecure Skill Coding Practices

Error
Location
SKILL.md:47
Finding

BigCommerce API Token Exposed Through Command-Line Arguments

Content
View full analysis
Remediation
View remediation

T08 · Insecure Dependencies

Warning
Location
SKILL.md:33
Finding

Unpinned Dependency Installation Bypasses System Package Protections

Content
View full analysis
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Error
Location
scripts/bc_optimizer.py:241
Finding

Unvalidated Agent-Generated HTML and Product Updates Are Written to the Storefront

Content
View full analysis
`, `
    `, `
  • ` tags. Focus on benefits, use cases, value proposition. No inline styles, no scripts. ``` ### Technical Analysis The executable code does not enforce the documented HTML restrictions. It reads an arbitrary JSON file, accepts the supplied product ID, and forwards supported fields directly to the BigCommerce API. No code-level controls verify: - That descriptions contain only `p`, `ul`, `li`, `strong`, and `em`. - That elements lack event-handler or other dangerous attributes. - That field values are strings of acceptable length. - That product IDs belong to the page previously fetched. - That every update corresponds to an authorized, reviewed product. - That the generated payload contains a meaningful and safe change. - That a human approved the final diff before a bulk write. An instruction such as “No inline styles, no scripts” is not a security boundary. Agent-generated output, malformed update files, or locally modified files can violate it. Product data used as generation context may also contain adversarial text intended to influence generated output. Whether unsafe markup exe ...[truncated 1918 chars]
Remediation
View remediation

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Rogue AgentSelf-Modification, Session Persistence
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Findings (5)

Missing User Warnings

High
Category
Not specified by scanner
Confidence
97% confidence
Finding

The skill is designed to autonomously update live BigCommerce product titles and descriptions page by page without pausing, but it does not present a prominent warning or confirmation requirement before performing those bulk writes. In a production ecommerce environment, accidental execution can cause widespread content corruption, SEO damage, brand inconsistency, and time-consuming rollback efforts.

Content

No source excerpt is available for this finding.

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
90% confidence
Finding

The skill performs file reads/writes and networked API calls but does not declare any explicit tool scope or permissions boundaries. That increases the chance of over-privileged execution, weak reviewability, and unintended access to local files or external services beyond what the user expects.

Content

No source excerpt is available for this finding.

Session Persistence

Medium
Category
Rogue Agent
Confidence
75% confidence
Finding

The skill persists execution state in progress.json and is explicitly designed to resume in-progress bulk modifications automatically. Persistent state itself is not inherently malicious, but in this context it can cause unintended continuation of prior write operations, especially if the environment is reused, the file is stale, or a later invocation occurs with ambiguous intent.

Content

Scanner excerpt · SKILL.md (reported line 6)May include surrounding context.

md
description: >
  Autonomous BigCommerce product content optimizer. Bulk-update, rewrite, optimize,
  or generate product titles and descriptions on a BigCommerce store. Triggers on
  "optimize my BigCommerce products", "rewrite product descriptions", "update product
  content", "improve my store listings", "bulk update BigCommerce", or when BigCommerce
  API credentials (store hash, API token) are provided for content work. Handles the
  full lifecycle: fetching products page-by-page, generating improved content, updating

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The trigger phrases are broad enough that ordinary requests like improving listings or providing BigCommerce credentials could activate an autonomous bulk-update workflow. In this skill's context, that is dangerous because activation can quickly lead to live product rewrites across many items without a narrow, explicit user command.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
60% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · scripts/bc_optimizer.py (reported line 35)May include surrounding context.

python
# ---------------------------------------------------------------------------
# Constants
# ---------------------------------------------------------------------------
BASE_URL = "https://api.bigcommerce.com/stores/{store_hash}/v3"
PROGRESS_FILE = "progress.json"
RATE_LIMIT_PAUSE = 3          # seconds to wait when rate-limited
MAX_RETRIES = 5               # max retries per request on rate limit / transient error

Static analysis

No suspicious patterns detected.