Back to skill

Security audit

Autonomous Product Funnel

Security checks for vulnerabilities and agentic risk

Overview

ClawScan rates this skill suspicious with medium confidence, mainly because scanners flagged instructions allowing autonomous decisions and coverage was partial.

Consider the skill’s permission for routine autonomous decisions before installing, and check whether that level of control fits your needs. Treat the review as incomplete: Skillspector scanned 6 of 7 components, reported one partial inspection due to a parser limit, and found several local references it could not resolve. AIG completed with no findings, but that does not fill the Skillspector coverage gaps.

Vulnerability Patterns
  • 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
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Vulnerability Patterns
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Findings (7)

Credential Access

High
Category
Privilege Escalation
Confidence
70% confidence
Finding

Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Content

Scanner excerpt · references/tracking-and-integrations.md (reported line 42)May include surrounding context.

md
- Use the same logical event identifier for deduplication when supported.
- Fire Purchase at verified purchase success, not on checkout click.
- Preserve browser attribution cookies/parameters according to current Meta guidance and consent policy.
- Keep access tokens server-side.
- Use Test Events or equivalent validation when access exists.

## GA4

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
75% 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.

Content

Scanner excerpt · SKILL.md (reported line 84)May include surrounding context.

md
- If the probabilities are near-even, or confidence is low, do not pick on intuition: apply the decision priority below and take the most reversible option. Escalate to the user only when the decision is irreversible, spends money, or makes legal claims.
- If `typesafe-ai` is unavailable, the agent acts as JEV and applies the decision priority below directly.

JEV may decide without asking again:

- target segment and primary use case;
- positioning and value proposition;

Autonomous Decision Making

Medium
Category
Excessive Agency
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.

Content

Scanner excerpt · SKILL.md (reported line 172)May include surrounding context.

md
5. continue with all other work;
6. list the exact manual action only in the final handoff unless immediate intervention is strictly required to avoid destructive or unsafe behavior.

Never ask the user to paste raw secrets, private keys, CAPI tokens, passwords, or recovery codes into chat. Ask for connector availability, secret-manager path, or environment variable name instead.

## Research and truthfulness

Autonomous Decision Making

Medium
Category
Excessive Agency
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.

Content

Scanner excerpt · references/intake-questionnaire.md (reported line 79)May include surrounding context.

md
- Warranty information
- Known limitations or defects that must be disclosed

Never ask the user to claim a certification without evidence.

## 5. Market and customer-facing language

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
75% 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.

Content

Scanner excerpt · references/intake-questionnaire.md (reported line 562)May include surrounding context.

md
End intake with one explicit authorization statement, translated into the user's language:

"After I submit this intake, JEV may make routine decisions within the authority and constraints above. The agent may continue through research, strategy, copy, implementation, payment, tracking, deployment, and QA without asking for ordinary approvals. If an external human-only action is unavailable, complete all surrounding work and list the exact remaining action in the final handoff."

Collect:

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
85% 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.

Content

Scanner excerpt · references/tracking-and-integrations.md (reported line 103)May include surrounding context.

md
- existing tracking gates;
- user matching policy.

Do not silently bypass consent logic to improve attribution.

## Tracking matrix template

Internal documentation is required to be in English without user opt-in.

Low
Category
Output Handling
Confidence
90% confidence
Finding

The instruction mandates English for internal documentation and bundled references without offering a choice or documenting a justification. The subsequent conversation-language policy accommodates user-facing communication but does not provide an exception to this documentation requirement.

Content

No source excerpt is available for this finding.