Back to skill

Security audit

meta-autonomous-deep-research

Security checks for vulnerabilities and agentic risk

Overview

This skill is a small, disclosed research-workflow prompt with a local learner log, but its claimed autonomous research capabilities are mostly aspirational rather than implemented.

Install only if you want a Chinese-language research workflow prompt with a simple persistent local learner log. Avoid passing secrets or sensitive task details as learner notes, and do not assume the advertised self-verification, orchestration, or full research automation exists unless you add or verify those supporting tools separately.

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
  • 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
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (5)

Tp4

High
Category
MCP Tool Poisoning
Confidence
96% confidence
Finding

The implementation only loads and updates a local JSON memory file with totals for operations, failures, and notes. While this could be a small supporting component for a broader meta-evolver loop, the supplied code chunk itself does not implement the declared high-level capabilities such as autonomous research, orchestration, self-verification, or self-reflection. Its primary purpose is persistence of simple telemetry/learning records, which is materially narrower than the declared description.

Content

No source excerpt is available for this finding.

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
94% confidence
Finding

The skill advertises autonomous research, self-reflection, and learner integration, and the analysis detected file-write capability, yet the manifest declares no explicit tool scope or permission boundary. In an agent ecosystem, undocumented write access undermines least-privilege controls and can let the skill persist data, alter artifacts, or create unintended side effects without operator awareness.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The natural-language content of the skill description and usage instructions is entirely in Chinese, and there is no indication that users may choose another language or that the skill is intentionally limited to a Chinese-speaking or region-specific context. This creates a language/locale policy concern because the skill implicitly enforces a specific language without 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

This markdown file is entirely written in Chinese and does not indicate that language selection is optional or region-specific. Under the policy, forcing a specific language without user opt-in is a natural-language policy violation.

Content

No source excerpt is available for this finding.

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
93% confidence
Finding

The script persists a caller-controlled note string into a long-lived JSON file without any notice, consent, retention control, or sanitization. In this skill context, the file is explicitly used to build a self-improving memory loop, so prompts, task details, or other sensitive operational metadata could be silently accumulated and later exposed to other components or users.

Content

No source excerpt is available for this finding.