Back to skill

Security audit

AI技能制造机(写/体检/升级技能包)

Security checks for vulnerabilities and agentic risk

Overview

ClawScan rates this skill suspicious because it found an unannounced change to persistent skill notes and could not fully inspect a referenced script; this is not a finding of malicious intent.

Consider the verdict suspicious and review the skill’s behavior before installing. In particular, account for the automatic persistent ledger update; one referenced script was only partially inspected, so the evidence is incomplete.

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

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 35)May include surrounding context.

md
| skill安检/体检/评分/质量/审计 | 安检 | scripts/skill-audit.js + references/audit-rubric.md |

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 44)May include surrounding context.

md
机检:`node scripts/skill-audit.js <目录>`(frontmatter/死链/孤儿文件/双源/体积/脚本语法/模糊纪律词)→ FAIL 清零才做语义九维。

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 77)May include surrounding context.

md
□ 本skill自身改动后 `node scripts/skill-audit.js <本目录>` exit 0 —— 贴汇总行

Incomplete referenced artifact analysis

High
Category
analysis-evasion
Confidence
100% confidence
Finding

Referenced artifact was not completely inspected

Content

Scanner excerpt · SKILL.md (reported line 91)May include surrounding context.

md
| scripts/skill-audit.js | 机器安检器(被检skill与本skill通用) |

The workflow automatically modifies the skill’s lessons ledger without warning the user about this persistent change.

Medium
Category
Output Handling
Confidence
92% confidence
Finding

The audit workflow instructs the agent to append lessons to the installed skill’s own persistent ledger without a confirmation gate, even though saving the audit report explicitly requires user approval. This expands a review into modification of reusable skill state; inaccurate or adversarial material incorporated as lessons could influence subsequent tasks.

Content

No source excerpt is available for this finding.

The template requires a Chinese skill name without user language opt-in.

Low
Category
Output Handling
Confidence
92% confidence
Finding

The placeholder <skill 名·中文> explicitly specifies Chinese for the generated skill name. Unlike merely writing the template's instructions in Chinese, this constrains generated output to a particular language without offering a choice or documenting a locale-specific justification.

Content

No source excerpt is available for this finding.