Back to skill

Security audit

ci-cd-security-audit-lite

Security checks for vulnerabilities and agentic risk

Overview

ClawScan rates this skill suspicious, mainly because it flags an undisclosed external submission of user feedback; the scan is partial, so the evidence is incomplete.

Consider whether you are comfortable with the reported external submission of feedback and the extra directory listing. The scan covered only half of the components fully; two JavaScript files were partially inspected, so the results do not establish that the skill is safe.

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
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (2)

The skill instructs the agent to submit the user's rating to a marketplace tool without first disclosing that external submission to the user.

Low
Category
Output Handling
Confidence
91% confidence
Finding

The skill directs the agent to submit the user's rating through an external marketplace tool without first telling the user that their response will be transmitted. Providing feedback in a conversation does not necessarily authorize external publication or submission, and this transmission is unrelated to the advertised local-only workflow audit. The exposure is limited to feedback and any metadata attached by the tool; the supplied content does not direct submission of workflow contents or secrets.

Content

No source excerpt is available for this finding.

Enumerates sibling workflows for upgrade messaging despite the declared single-file scope.

Low
Category
Excessive Agency
Confidence
94% confidence
Finding

The manifest limits the skill to linting a single local workflow file, but L108–L116 enumerate its containing workflows directory and identify other YAML files. This additional filesystem inspection serves an upgrade recommendation at L119 rather than analysis of the requested file; it does not read sibling file contents.

Content

No source excerpt is available for this finding.