Back to skill

Security audit

物理错误DNA

Security checks for vulnerabilities and agentic risk

Overview

ClawScan marked this skill suspicious with medium confidence because scanners reported excessive agency and a handover schema that may allow out-of-scope writebacks; intent is not established, and analysis was partial.

Consider the scope and handover findings before installing. Skillspector inspected all 14 listed components, but reported partial analysis because a local path-like reference in SKILL.md was missing; the evidence is therefore incomplete. The other scanner completed with no findings, which does not resolve the reported concerns.

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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (6)

The anxiety workflow directly performs concept rebuilding despite the manifest explicitly delegating it to another skill.

Medium
Category
Excessive Agency
Confidence
97% confidence
Finding

The manifest at L003 explicitly excludes concept rebuilding and routes it to the physics concept-intuition skill. However, L369 instructs this skill to explain concepts using three everyday analogies, followed by its own progressive exercises at L372–L376, rather than handing off the rebuilding task.

Content

No source excerpt is available for this finding.

Ae4

Medium
Category
analysis-evasion
Confidence
80% confidence
Finding

Suspicious Unicode normalization or mixed-script content

Content

No source excerpt is available for this finding.

The skill's mandatory remediation workflow includes concept reconstruction and solution-process coaching that the manifest explicitly delegates to other skills.

Medium
Category
Excessive Agency
Confidence
94% confidence
Finding

The manifest excludes concept reconstruction and solution-process guidance, requiring handoff to the physics concept-intuition and problem-solving coach skills. However, L373–L378 instruct this skill to perform analogy-based reconstruction and process-decomposition training, and L004 explicitly assigns Section Eight's execution to physics-error-dna. The referenced process-remediation method at L237–L242 extends through setting up and solving equations, rather than stopping at root-cause diagnosis or referral.

Content

No source excerpt is available for this finding.

Ae4

Medium
Category
analysis-evasion
Confidence
80% confidence
Finding

Suspicious Unicode normalization or mixed-script content

Content

No source excerpt is available for this finding.

The physics-specific outbound analysis contract permits writebacks for unrelated subjects.

Medium
Category
Excessive Agency
Confidence
92% confidence
Finding

The physics-specific schema permits deep-analysis writebacks with unrelated subjects, despite restricting the sender to the physics error-analysis skill. If a receiver relies on schema validation for subject authorization, a malformed or attacker-influenced handover could modify another subject’s error records; independent receiver-side authorization would mitigate this risk.

Content

No source excerpt is available for this finding.

The error-analysis skill directly performs solution coaching that its manifest explicitly delegates to another skill.

Medium
Category
Excessive Agency
Confidence
97% confidence
Finding

The manifest excludes solution-process guidance and delegates it to the physics problem coach. However, this section, explicitly assigned to physics-error-dna at L312, instructs the skill to correct diagrams in-session, require redrawing, and enter step-by-step guidance for severe errors. These instructions go beyond diagnosing and recording error causes into the excluded coaching workflow.

Content

No source excerpt is available for this finding.