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。
其他技能

perseus-specialist

Run all specialist deep-dive skills in parallel for comprehensive analysis

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-specialist

技能说明


name: perseus-specialist description: Run all specialist deep-dive skills in parallel for comprehensive analysis

Perseus Specialist (All Deep-Dive Skills)

Context & Authorization

IMPORTANT: This skill runs all specialist security analyses on the user's own codebase. This is defensive security testing to find vulnerabilities before attackers do.

Authorization: The user owns this codebase and has explicitly requested this comprehensive analysis.


Overview

This skill orchestrates ALL 8 specialist skills in parallel for maximum coverage and efficiency.

When to Use:

  • After /scan to get deep analysis on all detected components
  • When you want comprehensive specialist coverage without running /start
  • As a standalone deep-dive after initial assessment

Goal: Run all specialists simultaneously for thorough security analysis.

Engagement Mode Compatibility

ModeOrchestrator Behavior
PRODUCTION_SAFERun specialists with passive-first and low-rate verification constraints
STAGING_ACTIVEAllow active specialist checks with throttling
LAB_FULLAllow broad dynamic specialist verification in isolated lab
LAB_RED_TEAMAllow chain-based specialist simulation with strict kill-switches

Safety Gates (Required)

  1. Read deliverables/engagement_profile.md before launching specialists.
  2. Default to PRODUCTION_SAFE if engagement mode is not available.
  3. Propagate mode and rate limits to each specialist task.
  4. Stop all specialists if any run reports ABORTED-SAFETY.

Specialists Included

SkillCoverageOutput
perseus-apiOWASP API Top 10, GraphQL, WebSocketapi_security_analysis.md
perseus-injectionNoSQL, LDAP, XPath, SSTI, Commandinjection_deep_analysis.md
perseus-cryptoJWT, Hashing, Encryption, Secretscrypto_security_analysis.md
perseus-supply-chainCVEs, Dependencies, Licensessupply_chain_analysis.md
perseus-filePath Traversal, Upload, XXEfile_security_analysis.md
perseus-logicRace Conditions, Business Logicbusiness_logic_analysis.md
perseus-clientDOM XSS, Prototype Pollutionclient_side_analysis.md
perseus-configHeaders, CORS, Cookies, TLSconfig_security_analysis.md

Execution Instructions

Step 0: Mode & Scope Alignment

  • Load mode/scope/limits from deliverables/engagement_profile.md.
  • Respect deliverables/verification_scope.md when present.
  • Announce mode before launching specialists.

Step 1: Announce Start

"Running all Perseus specialist skills in parallel..."
"This provides deep-dive analysis across 8 security domains."

Step 2: Launch All Specialists (Parallel or Sequential)

Adapt execution to your runtime:

  • Mode A (Subagents Available): Launch 8 parallel tasks using a single message with subagent tool calls (Task, Agent, or invoke_subagent):
    1. Task: "Run API security specialist" -> Skill: perseus-api
    2. Task: "Run injection specialist" -> Skill: perseus-injection
    3. Task: "Run crypto specialist" -> Skill: perseus-crypto
    4. Task: "Run supply chain specialist" -> Skill: perseus-supply-chain
    5. Task: "Run file security specialist" -> Skill: perseus-file
    6. Task: "Run business logic specialist" -> Skill: perseus-logic
    7. Task: "Run client-side specialist" -> Skill: perseus-client
    8. Task: "Run config specialist" -> Skill: perseus-config
  • Mode B (Single-Agent Direct Execution — UNIVERSAL): If subagents are not available, do not attempt to call non-existent tools. Systematically run the checks for each specialist sequentially in the current session and write deliverables directly.

Step 3: Wait for Completion

Wait for all 8 specialists to complete their analysis.

Step 4: Summarize Results

"Specialist analysis complete!"

Summary:
- API Security: X findings
- Injection: X findings
- Cryptography: X findings
- Supply Chain: X findings
- File Security: X findings
- Business Logic: X findings
- Client-Side: X findings
- Configuration: X findings

Total: X findings across 8 domains

"All reports saved to deliverables/"

Output Structure

After completion, deliverables/ will contain:

deliverables/
├── api_security_analysis.md
├── injection_deep_analysis.md
├── crypto_security_analysis.md
├── supply_chain_analysis.md
├── file_security_analysis.md
├── business_logic_analysis.md
├── client_side_analysis.md
└── config_security_analysis.md

When to Use Each Specialist Individually

If You NeedRun
Only API analysis/api
Only injection deep-dive/injection
Only crypto audit/crypto
Only dependency check/supply-chain
Only file/upload security/file
Only business logic/logic
Only client-side/client
Only config hardening/config
All of the above/specialist

Integration with Core Skills

Recommended Flow:

/scan              → Map attack surface
    ↓
/specialist        → Deep-dive all domains (this skill)
    ↓
/audit             → Core vulnerability analysis
    ↓
/exploit           → Verify findings
    ↓
/report            → Generate final report

Or simply:

/start             → Runs everything automatically

Quick Reference

CommandWhat It Does
/specialistAll 8 specialists in parallel
/startFull assessment (includes specialists)
/apiAPI security only
/injectionInjection analysis only
/cryptoCryptography only
/supply-chainDependencies only
/fileFile security only
/logicBusiness logic only
/clientClient-side only
/configConfiguration only