Back to skill

Security audit

Solo Ecommerce Agent

Security checks for vulnerabilities and agentic risk

Overview

The skill is not clearly malicious, but it encourages unattended live-store automation that can publish products, message customers, and process shipments without adequate safeguards.

Review carefully before installing on any real store account. Do not enable cron or live publishing, customer replies, shipping, refunds, browser automation, or API keys until you have explicit approval gates, dry-run testing, audit logs, scoped credentials, and a documented way to disable every scheduled task.

Vulnerability Patterns
  • System PersistenceInstalls backdoors, hooks, services, or scheduled tasks that survive the run
  • 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)

T06 · System Persistence

Error
Location
SKILL.md:40
Finding

Skill-load instructions request persistent scheduled execution

Content
View full analysis

Vulnerability Details

File Location: SKILL.md, lines 40-48
Vulnerability Type: Persistent scheduled-task registration
Risk Level: High

Vulnerable Snippet

The following is an English translation of the relevant source excerpt:

markdown
### Step 2: Enable scheduled tasks

After the skill is loaded, create the following cron tasks:

| Task | Schedule | Purpose |
| Product scanning | Every hour | Analyze bestseller lists |
| Customer service | Every 5 minutes | Automatically process messages |
| Order processing | Every 10 minutes | Automatically ship and record logistics |
| Daily report | Every day at 23:00 | Send daily operating data |

Technical Analysis

The instruction associates loading the skill with creating four recurring cron tasks. Scheduled tasks continue to execute after the initiating interaction has ended and therefore establish cross-session persistence.

This is particularly sensitive because the described tasks include customer-message handling, product publishing, order processing, shipping, and logistics updates. Once the corresponding integrations are implemented and enabled, repeated execution could cause externally visible commercial actions without contemporaneous review.

The current Python scripts do not directly install cron entries, and their platform integrations are placeholders. Exploitation therefore depends on the hosting agent following the skill instruction and registering the schedules. No credential theft, data exfiltration, destructive payload, or remote code retrieval was identified in the reviewed package.

Attack Path

  1. A user installs or loads the skill.
  2. The agent processes the instruction in SKILL.md to create recurring cron tasks.
  3. The tasks remain registered after the original session ends.
  4. The scheduler repeatedly invokes product, customer-service, order, and reporting scripts.
  5. If platform integrations and automation fl ...[truncated 897 chars]
Remediation
View remediation

Remediation Suggestions

  1. Do not register scheduled tasks automatically when the skill is loaded.
  2. Require explicit, informed user approval before creating each schedule.
  3. Display the exact command, interpreter, script path, frequency, execution account, and affected store before confirmation.
  4. Default transactional functions such as publishing and shipping to disabled and human-review-required states.
  5. Use a dedicated least-privilege service account with access only to the required data directory and platform operations.
  6. Add execution limits, idempotency controls, transaction caps, and emergency disable controls.
  7. Document how to list, disable, and remove every task created by the skill.
  8. Require renewed authorization before enabling integrations that can publish products, send messages, approve refunds, or confirm shipments.
  9. Record schedule creation, modification, execution, and removal in an auditable log.

T09 · Insecure Skill Coding Practices

Note
Location
scripts/customer_service.py:38
Finding

Undefined configuration loader causes the customer-service task to fail

Content
View full analysis

Vulnerability Details

File Location: scripts/customer_service.py, lines 38-40
Vulnerability Type: Unhandled undefined function in a scheduled entry point
Risk Level: Low

Vulnerable Snippet

python
def main():
    cfg = load_config()
    if not cfg or not cfg.get("enabled"):

No load_config function is defined or imported anywhere in scripts/customer_service.py.

Technical Analysis

Calling an undefined Python name raises NameError. Every invocation of main() reaches this call before configuration validation or feature-state checks occur. Consequently, both enabled and disabled configurations fail before the script can safely decide whether customer-service processing should run.

When invoked by the documented five-minute schedule, this defect creates a persistent availability failure and may generate repeated scheduler errors. It also prevents the script from honoring its documented disable control.

This is not a direct confidentiality or privilege-escalation vulnerability. Its security relevance is operational: an automated customer-service control fails closed through an unhandled exception, potentially concealing urgent customer or dispute messages if operators assume the scheduled task is functioning.

Attack Path

  1. The customer-service script is invoked manually or by the documented recurring schedule.
  2. Python enters main().
  3. The interpreter evaluates load_config().
  4. Because that name is neither defined nor imported, Python raises NameError.
  5. The process terminates before checking enabled or customer-service automation settings.
  6. Repeated scheduled invocations continue to fail until the source is corrected.

No attacker-supplied input is required to trigger the failure.

Impact Assessment

The issue does not grant additional privileges or code execution. Its scope is the availability and reliability of the customer-service component. Automa ...[truncated 211 chars]

Remediation
View remediation

Remediation Suggestions

  1. Implement a configuration loader before main():

    python
    def load_config():
        if not CONFIG_FILE.exists():
            return None
        try:
            return json.loads(CONFIG_FILE.read_text(encoding="utf-8"))
        except (OSError, json.JSONDecodeError):
            return None
    
  2. Log configuration errors without exposing secrets or full sensitive configuration contents.

  3. Return a nonzero exit status for scheduler-visible failures.

  4. Add tests covering missing, malformed, disabled, and enabled configurations.

  5. Add a smoke test that invokes every declared script entry point.

  6. Configure scheduler monitoring to alert after repeated failures instead of silently retrying indefinitely.

T09 · Insecure Skill Coding Practices

Warning
Location
scripts/init_agent.py:12
Finding

Conflicting data directories and configuration schemas bypass the intended operating controls

Content
View full analysis

Vulnerability Details

File Location: scripts/init_agent.py, lines 12 and 47-64; scripts/daily_report.py, line 12; operational scripts at lines 6-8
Vulnerability Type: Inconsistent security configuration and state storage
Risk Level: Medium

Vulnerable Snippets

The initializer and report use a top-level home directory:

python
# scripts/init_agent.py
DATA_DIR = Path.home() / "solo-ecommerce-data"
python
# scripts/daily_report.py
DATA_DIR = Path.home() / "solo-ecommerce-data"

Operational scripts use a different directory:

python
# scripts/product_scanner.py
DATA_DIR = Path.home() / ".qclaw" / "solo-ecommerce-data"
DATA_DIR.mkdir(parents=True, exist_ok=True)
CONFIG_FILE = DATA_DIR / "config.json"
python
# scripts/product_publisher.py
DATA_DIR = Path.home() / ".qclaw" / "solo-ecommerce-data"
CONFIG_FILE = DATA_DIR / "config.json"
python
# scripts/customer_service.py
DATA_DIR  = Path.home() / ".qclaw" / "solo-ecommerce-data"
CONFIG_FILE = DATA_DIR / "config.json"
python
# scripts/order_processor.py
DATA_DIR  = Path.home() / ".qclaw" / "solo-ecommerce-data"
CONFIG_FILE = DATA_DIR / "config.json"

The initializer writes this schema:

python
config = {
    "version": "1.0.0",
    "created": get_timestamp(),
    "platforms": platforms,
    "automation_scope": {
        "auto_publish": automation_config.get("auto_publish", False),
        "auto_pricing": automation_config.get("auto_pricing", False),
        "auto_reply": automation_config.get("auto_reply", True),
        "auto_ship": automation_config.get("auto_ship", False),
        "complaint_to_manual": True,
        "refund_threshold": automation_config.get("refund_threshold", 50)
    },
    "risk_controls": {
        "publish_needs_review": automation_config.get("publish_needs_review", True),
        "ship_needs_confirm": automation_config.g
...[truncated 2886 chars]
Remediation
View remediation

Remediation Suggestions

  1. Define a single canonical data directory in a shared module and import it from every script.
  2. Align the implementation with the documented ~/.qclaw/solo-ecommerce-data location, or update all documentation and code together if another location is selected.
  3. Define one versioned configuration schema containing the same keys consumed by operational modules.
  4. Validate configuration with a formal schema before any scheduled or transactional operation.
  5. Fail safely on unknown, missing, or conflicting keys rather than silently treating them as disabled or enabled.
  6. Ensure review controls are checked immediately before product publication, shipping, refunds, or other irreversible actions.
  7. Provide a one-time migration that detects both legacy directories, reports conflicts, and requires user approval before merging or deleting data.
  8. Set restrictive directory and file permissions because future files may contain customer, order, API, or store information.
  9. Add integration tests that run initialization followed by every operational script and the daily report.
  10. Include the effective configuration path and schema version in sanitized audit logs.
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 (18)

Missing User Warnings

High
Category
Not specified by scanner
Confidence
95% confidence
Finding

The README explicitly promotes fully automated listing, customer service, shipping, and refund review with no warnings, approval gates, or discussion of data-handling and irreversible business consequences. In an e-commerce agent, these actions can publish incorrect products, mishandle customer communications, approve fraudulent refunds, or trigger shipping/order changes affecting real customers and business accounts.

Content

No source excerpt is available for this finding.

Missing User Warnings

High
Category
Not specified by scanner
Confidence
95% confidence
Finding

The sample configuration enables automated publishing, customer service, and order processing by default, which encourages unsafe deployment without informed consent or staged rollout. Because this is copy-pasteable setup guidance, users may activate real-world business automation immediately, increasing the chance of unauthorized listings, bad customer responses, incorrect fulfillment, or harmful refund decisions.

Content

No source excerpt is available for this finding.

Missing User Warnings

High
Category
Not specified by scanner
Confidence
96% confidence
Finding

The skill advertises fully automated publishing, customer service, shipping, refunds, and daily operations with '无需人工干预' and frequent scheduled execution, but it does not present prominent warnings about business, legal, privacy, or irreversible operational consequences. In this context, the omission is dangerous because the agent is positioned to act on live store inventory, customer communications, and order fulfillment, which can cause unauthorized transactions, privacy mishandling, reputational harm, or financial loss if misconfigured or triggered unexpectedly.

Content

No source excerpt is available for this finding.

Missing User Warnings

High
Category
Not specified by scanner
Confidence
98% confidence
Finding

The sample configuration enables automatic publishing, auto-reply, and auto-shipping with review disabled, while providing no caution that these actions can immediately affect customers, orders, and storefront state. This is especially risky because the config is presented as a quick-start template, making it likely users will copy unsafe defaults that can trigger unintended listings, erroneous shipments, incorrect refund handling, or disclosure/misuse of customer data.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
83% confidence
Finding

The README is entirely presented in Chinese and describes the skill as if that is the required operating language, but it does not offer alternative language support or explain a necessary locale limitation. This can violate language or locale policy when a skill forces a specific language without user opt-in.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

The skill title, instructions, and trigger examples are entirely in Chinese and the interaction examples presume Chinese-language use, with no indication that other languages are supported or that the user can choose a locale. This creates a language-policy issue because the skill effectively forces a specific language without opt-in.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The description claims 'full-platform' and 'fully automated' ecommerce operations across product selection, listing, customer service, and order handling without any stated scope limits, approval gates, or safety constraints. In an agent skill, this can encourage overbroad invocation and high-risk autonomous actions affecting storefronts, customers, and transactions, especially given the declared browser automation dependency.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
84% confidence
Finding

The manifest description is only in Chinese and does not indicate any language choice or localization support. This can mislead or exclude users and reviewers who cannot read the description, reducing informed consent and oversight for a skill that appears capable of broad automation across commercial workflows.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The guide gives step-by-step instructions for automating high-impact seller actions such as publishing products and confirming shipment, but it omits safeguards like operator approval, dry-run validation, or warnings about irreversible business effects. In the context of browser automation over already-authenticated seller accounts, an agent could perform unintended listings or fulfillment actions immediately, causing financial, compliance, or customer-service harm.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

The customer-service workflow instructs the agent to select a conversation, enter a reply, and send it without any requirement for human review, privacy checks, or accuracy validation. Because this operates in live buyer conversations on logged-in merchant systems, automated transmission could leak sensitive data, send misleading statements, or create legal and reputational issues.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The API section encourages use of platform APIs and mentions keys, but it provides no instructions for secure credential storage, least-privilege scoping, rotation, or protecting store data sent through external integrations. This omission is risky because users may embed secrets insecurely or expose merchant data while implementing automation at scale.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The file’s top-level description states the skill is a customer-service engine for automatically replying to buyers, and all user-facing replies are hard-coded in Chinese. There is no indication of language selection, user opt-in, or a documented region-specific constraint, which creates a natural-language locale policy concern.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
98% confidence
Finding

The file's docstrings, prompts, status messages, and usage instructions are all hard-coded in Chinese, beginning with the top-level description and continuing throughout the interactive flow. Under the policy, forcing a specific language without user opt-in is a natural-language policy violation unless the locale restriction is explicitly justified.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

This code's natural-language interface is entirely in Chinese, including the module description and all user-facing log messages. The file does not offer any language selection, opt-in, or justification that the skill is limited to a Chinese-speaking or region-specific environment, which creates a locale-policy concern.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The module docstring is entirely in Chinese and presents the skill as a general-purpose product publishing engine, but there is no indication that Chinese is optional or that the skill is limited to a Chinese-speaking or region-specific context. This creates a natural-language locale policy issue because the skill implicitly fixes the interaction language without offering user choice or documenting a justified constraint.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The module docstring and all user-facing log/error messages are written in Chinese, which imposes a specific language on users. The file does not offer any language/locale selection or explain that the skill is intended only for a Chinese-speaking or region-specific context.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
79% confidence
Finding

All instructional content and example user utterances are presented only in Chinese, and the file does not state that the skill is region-specific or offer any language/locale option. Per SQP-3, forcing a specific language without user opt-in can be a natural-language policy issue unless the locale constraint is clearly justified.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
89% confidence
Finding

This code embeds user-facing descriptions and console/report output in Chinese throughout the file, starting with the module docstring. The policy requires flagging language or locale constraints when the skill forces a specific language without user opt-in, and no alternative language option or justification is provided here.

Content

No source excerpt is available for this finding.