跳至正文
OpenAgentSkill
浏览已开放 · 目录记录日期 2026-11-21 · 自动更新和账户服务暂时关闭。
安全

perseus-report

Use when generating the final executive security report (Phase 4)

kaivyy · 已有静态检查记录

安装与使用

运行命令前先阅读来源与依赖说明。复制命令不会自动安装或执行。

npx skills add kaivyy/perseus --skill perseus-report

技能说明


name: perseus-report description: Use when generating the final executive security report (Phase 4)

Perseus Report (Phase 4)

Context & Authorization

IMPORTANT: This skill generates a professional security report for the user's own codebase. The report is intended to help developers and stakeholders understand and remediate security issues in their own software.


Multi-Language Support

Report generation adapts remediation guidance based on detected language/framework:

LanguageFrameworks
JavaScript/TypeScriptExpress, Fastify, Next.js, Nest.js
GoGin, Echo, Fiber, Chi
PHPLaravel, Symfony, Slim
PythonFastAPI, Django, Flask
RustActix-web, Axum, Rocket
JavaSpring Boot, Quarkus
RubyRails, Sinatra
C#ASP.NET Core

Overview

This skill executes the Reporting Phase of the Perseus framework. It synthesizes findings from Scan, Audit, Exploit, and Specialist phases into a professional executive report.

Goal: Communicate verified risks effectively to stakeholders and drive remediation.

Methodology:

  1. Collect: Gather all deliverables from previous phases.
  2. Verify: Prioritize verified exploits over theoretical risks.
  3. Contextualize: Explain business impact, not just technical flaws.
  4. Remediate: Provide language-specific, actionable fixes and unified git diff patches for each finding.
  5. Standardize: Output both human-readable Markdown and machine-readable SARIF formats.

Agent Execution Strategy (Dual-Mode)

When executing this skill, adapt to the runtime environment:

  • Direct Synthesis (Universal): Read deliverables from deliverables/, aggregate findings, calculate CVSS scores and CWE classifications, and write the final reports directly in the current session. No subagent spawning is required for report generation.

Execution Instructions

Step 1: Collect All Deliverables

Read all files in deliverables/:

Core Phase Deliverables:

  • engagement_profile.md (Mode, scope, constraints)
  • code_analysis_deliverable.md (Scan)
  • All *_analysis.md files (Audit)
  • exploitation_report.md (Exploit)
  • verification_scope.md (Verification limits and approved window)

Specialist Deliverables (if present):

  • api_security_analysis.md (API + GraphQL + WebSocket + OAuth)
  • injection_deep_analysis.md (NoSQL + SSTI + Log4j)
  • crypto_security_analysis.md (JWT + Hashing + Encryption)
  • supply_chain_analysis.md (CVEs + Typosquatting)
  • file_security_analysis.md (Path Traversal + XXE + Zip Slip)
  • business_logic_analysis.md (Race Conditions + AI Security)
  • client_side_analysis.md (React/Vue/Angular + SSR)
  • config_security_analysis.md (Docker + CI/CD + Cloud + K8s)

Step 2: Calculate Risk Metrics

Severity Scoring (CVSS-inspired):

SeverityCriteriaScore
CriticalRCE, Auth Bypass, SQLi with data access, Admin takeover, Container escape, AI prompt injection leading to RCE9.0-10.0
HighStored XSS, SSRF to internal, Privilege Escalation, Sensitive data exposure, CI/CD injection, Public cloud storage7.0-8.9
MediumReflected XSS, CSRF, Information disclosure, Missing security headers, Outdated dependencies, Docker misconfig4.0-6.9
LowMinor info leak, Best practice violations, Verbose errors, License issues0.1-3.9

Exploitability Factor:

  • VERIFIED: Multiply by 1.0 (confirmed exploitable)
  • POTENTIAL: Multiply by 0.7 (likely exploitable)
  • THEORETICAL: Multiply by 0.4 (needs specific conditions)

Confidence Factor:

  • High: Multiply by 1.0
  • Medium: Multiply by 0.75
  • Low: Multiply by 0.5

Verification Context Factor:

  • VERIFIED in PRODUCTION_SAFE: Multiply by 1.0
  • VERIFIED in STAGING_ACTIVE: Multiply by 0.9
  • VERIFIED in LAB_FULL: Multiply by 0.85
  • VERIFIED in LAB_RED_TEAM: Multiply by 0.8
  • POTENTIAL-PROD-BLOCKED: Multiply by 0.7

Step 3: Generate Report

Create deliverables/SECURITY_REPORT.md using this structure:

# Security Assessment Report

**Project:** [Project Name]
**Assessment Date:** [Date]
**Methodology:** Perseus Security Framework v2.0
**Scope:** [Repository/Application name]
**Engagement Mode:** [PRODUCTION_SAFE/STAGING_ACTIVE/LAB_FULL/LAB_RED_TEAM]

---

## Executive Summary

### Assessment Status
[Complete/Partial] - [X] phases completed, [Y] specialists run

### Technologies Analyzed
| Category | Detected |
|----------|----------|
| Language | [e.g., TypeScript, Go, Python] |
| Framework | [e.g., Next.js 14, Gin, FastAPI] |
| Database | [e.g., PostgreSQL, MongoDB, Redis] |
| Infrastructure | [e.g., Docker, Kubernetes, AWS] |
| CI/CD | [e.g., GitHub Actions, GitLab CI] |
| AI/LLM | [e.g., OpenAI, Anthropic, LangChain] |

### Risk Overview
| Severity | Verified | Potential | Total |
|----------|----------|-----------|-------|
| Critical | X | Y | Z |
| High | X | Y | Z |
| Medium | X | Y | Z |
| Low | X | Y | Z |

### Verification Coverage
| Category | Count |
|----------|-------|
| Verified | X |
| Failed Verification | Y |
| Potential (Prod Blocked) | Z |
| Aborted (Safety Kill-Switch) | W |

### Key Findings
1. **[Most Critical Finding]** - Brief description and impact
2. **[Second Critical Finding]** - Brief description and impact
3. **[Third Critical Finding]** - Brief description and impact

### Business Impact
- **Data Breach Risk:** [High/Medium/Low] - [Explanation]
- **Service Disruption:** [High/Medium/Low] - [Explanation]
- **Compliance Impact:** [Regulations affected: GDPR, PCI-DSS, HIPAA, SOC2, etc.]
- **Reputation Risk:** [High/Medium/Low] - [Explanation]

### Top 3 Recommendations
1. [Highest priority fix]
2. [Second priority fix]
3. [Third priority fix]

---

## Attack Surface Summary

### Technology Stack
- **Language:** [e.g., TypeScript]
- **Framework:** [e.g., Next.js 14 (App Router)]
- **Database:** [e.g., MongoDB, PostgreSQL]
- **Authentication:**
阅读完整来源 ↗