Back to skill

Security audit

成片参考分析

Security checks for vulnerabilities and agentic risk

Overview

ClawScan rated the video-reference skill suspicious with medium confidence because scanners found risks that exported references may include source names, dialogue, or timestamps despite stated restrictions.

Consider the risk that an exported reference could contain source names, dialogue, or timestamps if source checks are incomplete or a nested timestamp target is used. The scanner inspected all nine listed components, but reference resolution was partial because some local paths mentioned in SKILL.md were not found in the bundle, so the evidence is incomplete. Review exported material before reusing it.

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
  • Behavioral ASTexec() Call, eval() Call, Dynamic Import
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • 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
Findings (4)

Skill declares no tool scope ('permissions' or 'allowed-tools') but code capabilities were detected: file_read, file_write, shell.

Medium
Category
MCP Least Privilege
Confidence
70% confidence
Finding

Without declared permissions the skill's intent is opaque and cannot be validated.

Content

No source excerpt is available for this finding.

subprocess module call

Medium
Category
Dangerous Code Execution
Confidence
70% confidence
Finding

subprocess module calls execute external commands. Without careful input validation, this enables command injection.

Content

Scanner excerpt · scripts/lib.py (reported line 15)May include surrounding context.

python
def run_cmd(cmd, **kwargs):
    """Run a command list and return the CompletedProcess (stdout/stderr captured)."""
    return subprocess.run(
        [str(part) for part in cmd], capture_output=True, text=True,
        encoding="utf-8", errors="replace", **kwargs,
    )

Export can succeed without valid source material for the promised name/dialogue exclusion.

Medium
Category
Excessive Agency
Confidence
96% confidence
Finding

Missing or unusable ASR, missing research names, and a source-video identity mismatch generate warnings rather than blocking export. With empty manually supplied entities and an incomplete or unrelated corpus, source names or dialogue in method text can pass both scans and enter a reference intended for reuse in other projects; this is a local information-disclosure risk, not demonstrated network exfiltration.

Content

No source excerpt is available for this finding.

Method targets bypass the documented restriction that absolute source timestamps stay local.

Medium
Category
Excessive Agency
Confidence
92% confidence
Finding

The comment at L251 says the absolute s value stays local, but the redaction applies only when the target path exactly equals derived.first_original_at. A target using derived.first_original_at.s is resolved directly and exported unchanged, contradicting that stated export boundary.

Content

No source excerpt is available for this finding.