技能 mantis-dedupe
📦
mantis-dedupe
内容修订 r1 高风险 ⚙️ 外部命令📁 文件系统访问
Consolidate Duplicate Security Findings
Repeated security findings increase review effort and can obscure regressions. This skill guides consolidation while preserving provenance, active regression candidates, and transaction history.
支持: Claude Codex Code(CC)
安装前请停下并请求确认。
请审查安装计划，并在更改文件前获得用户明确同意。
使用我的智能体安装
将此请求复制给你的智能体，其中包含权威 Skill 页面和清单。
智能体请求
Review the Skillstore skill "mantis-dedupe" from https://skillstore.io/skills/google-mantis-dedupe.md and its manifest at https://skillstore.io/api/skills/google-mantis-dedupe/manifest. Verify the artifact. Stop and obtain explicit user consent before installing or changing files.你的智能体仍应展示其计划，并按安全策略请求必要确认。
Agent 可读资源
当 AI Agent、爬虫或脚本需要干净上下文，而不是读取完整页面时，请使用这些链接。
测试它
正在使用“mantis-dedupe”。 Two findings describe the same flaw at the same line and share discovery provenance.
预期结果:
- One primary finding retains its identifier and the highest reported severity.
- Descriptions, code paths, and history are consolidated.
- The duplicate is staged with linkage to the primary finding, and its move is logged.
正在使用“mantis-dedupe”。 A new finding matches a resolved archived issue, but their discovery snapshots differ.
预期结果:
- The new finding remains active.
- A possible duplicate hint references the archived finding.
- History records a possible regression; the finding is not moved to staging.
正在使用“mantis-dedupe”。 Two findings have matching signatures and titles but identify different lines in the same file.
预期结果:
- The findings remain separate because their line-inclusive locations differ.
- Signature agreement alone does not justify a duplicate status or staged removal.
安全审计
高风险v1 • 2026/10/4 打开版本化报告
The 160 static alerts describe Markdown notation, legitimate workspace operations, and defensive snapshot checks rather than the alleged execution or reconnaissance threats. Semantic review found unsafe helper reuse and potential disclosure of sensitive vulnerability summaries through external embeddings. No evidence found of malicious intent or instructions to manipulate this audit.
1
已扫描文件
372
分析行数
0
复核项
0
已忽略误报
已确认安全问题 (2)
高
Version Comment Does Not Authenticate Reused Helpers
The skill says, "Only reuse it when the marker matches." A modified helper can retain the expected comment and execute unauthorized code.
The instructions explicitly authorize helper reuse based on a version comment and later execute it. No content-integrity verification is specified.
中
External Embeddings May Disclose Vulnerability Details
The semantic fallback embeds root-cause summaries using a configured provider, defaulting to Vertex AI. These summaries include vulnerability mechanisms and dataflow without a disclosure-approval requirement.
The instructions specify sensitive summary fields and an external default embedding provider. Actual transmission depends on configured integration; no evidence found of malicious exfiltration intent.
风险因素
⚙️ 外部命令 (50)
SKILL.md:19 SKILL.md:23 SKILL.md:29 SKILL.md:30 SKILL.md:31 SKILL.md:33 SKILL.md:35 SKILL.md:36 SKILL.md:37 SKILL.md:38 SKILL.md:39 SKILL.md:40 SKILL.md:41 SKILL.md:42 SKILL.md:45 SKILL.md:47 SKILL.md:51 SKILL.md:52 SKILL.md:58-97 SKILL.md:97-100 SKILL.md:100 SKILL.md:101-103 SKILL.md:103-104 SKILL.md:104-105 SKILL.md:105-107 SKILL.md:107-108 SKILL.md:108-110 SKILL.md:110 SKILL.md:111 SKILL.md:123-125 SKILL.md:125-128 SKILL.md:128-129 SKILL.md:129-133 SKILL.md:133-136 SKILL.md:136-137 SKILL.md:137-138 SKILL.md:138-139 SKILL.md:139 SKILL.md:143-148 SKILL.md:148 SKILL.md:149 SKILL.md:150 SKILL.md:151 SKILL.md:152-153 SKILL.md:153-154 SKILL.md:154-155 SKILL.md:155 SKILL.md:156-157 SKILL.md:157 SKILL.md:161-162
审计者: codex
分享与引用此报告
分享版本化评估报告、中性徽章、嵌入卡片和引用。Skillstore 只报告证据，不判断该 Skill 是否安全。
复制报告链接
https://skillstore.io/skills/google-mantis-dedupe/audits/1?utm_source=security_passport&utm_medium=share&utm_campaign=versioned_reportMarkdown 徽章
[](https://skillstore.io/skills/google-mantis-dedupe?utm_source=security_passport_badge)HTML 徽章
<a href="https://skillstore.io/skills/google-mantis-dedupe?utm_source=security_passport_badge"><img src="https://skillstore.io/badges/skills/google-mantis-dedupe/security.svg" alt="Skillstore security assessment" loading="lazy"></a>嵌入卡片
<iframe src="https://skillstore.io/embed/skills/google-mantis-dedupe.html" title="Skillstore Security Assessment" sandbox="allow-popups allow-popups-to-escape-sandbox" loading="lazy" referrerpolicy="no-referrer" width="420" height="180"></iframe>学术引用 (APA · BibTeX · CFF)
APA 引用
google. (2026). mantis-dedupe security audit report (audit version 1) [Author version unspecified]. Skillstore. https://skillstore.io/skills/google-mantis-dedupe/audits/1BibTeX 引用
@techreport{google-google-mantis-dedupe-2026,
author = {google},
title = {mantis-dedupe security audit report (audit version 1)},
institution = {Skillstore},
year = {2026},
number = {1},
url = {https://skillstore.io/skills/google-mantis-dedupe/audits/1},
note = {Author version unspecified}
}CITATION.cff
cff-version: 1.2.0
message: "If you use this Skill, cite its author and this versioned security audit report."
title: "mantis-dedupe security audit report (audit version 1)"
version: "unspecified"
type: report
authors:
- name: "google"
date-released: "2026-10-04"
url: "https://skillstore.io/skills/google-mantis-dedupe/audits/1"
identifiers:
- type: other
value: "skillstore:google-mantis-dedupe:audit:1"
description: "Skillstore immutable audit report identifier"
Skillstore 评分
评分依据 证据置信度: 中 55
架构
85
可维护性
87
内容
65
社区
91
规范符合性
你能构建什么
Prepare a Security Review Queue
Consolidate repeated reports at identical locations while retaining severity, supporting details, and review history.
Preserve Regression Candidates Across Passes
Compare new findings with archived records and retain possible regressions when discovery snapshots differ.
Maintain Traceable Finding Pipelines
Process large queues with grouped comparisons, provenance checks, staged duplicates, and transaction records.
试试这些提示
Inspect a Small Finding Queue
Review the existing findings and identify duplicate candidates. Explain the matching evidence before changing any files.
Consolidate Current-Batch Duplicates
Consolidate findings with identical line-inclusive locations and matching titles. Preserve provenance and history, stage duplicates, and summarize changes.
Compare Findings Across Snapshots
Compare current and archived findings using pairwise discovery provenance. Keep unmatched findings active and annotate regressions against resolved archived issues.
Process a Large Queue Safely
Group findings by component, verify helpers before execution, and apply snapshot-gated matching. Require approval before external embeddings and summarize staged moves.
最佳实践
- Back up workspace state and review matching evidence before applying consolidation.
- Verify helper contents independently before execution; a version comment does not establish trust.
- Approve external embedding use explicitly and redact sensitive vulnerability details before transmission.
避免
- Marking findings as duplicates based only on signatures, without line-inclusive location agreement.
- Suppressing potential regressions or merging findings from different discovery versions.
- Executing existing helpers solely because their version comments match.
常见问题
Which AI tools can use this skill?
The marketplace lists Claude, Codex, and Claude Code. Workspace file access and command execution are needed to apply merges.
Does this skill discover new vulnerabilities?
No. It consolidates existing findings and identifies possible regressions during archive comparisons. Initial auditing and patch generation are outside its scope.
Are duplicate files permanently deleted?
No. Duplicates receive status and linkage metadata, move into a staging directory, and generate transaction records.
What happens when discovery snapshots differ?
Snapshot-gated archive matches remain active with a possible duplicate hint. Resolved archived matches receive possible regression notes.
Does deduplication require an external service?
Exact matching does not inherently require external embeddings. Semantic fallback requires configured embedding infrastructure and defaults to a Vertex AI model.
Is an existing helper safe when its version matches?
No. A matching comment does not verify script contents. Independently inspect or regenerate helpers before execution.