skilld
All skills
simota avatar

/probe

@e307415
by shingo imotasimota/agent-skills85 stars
15

Integrating OWASP ZAP/Burp Suite/Nuclei, planning penetration tests, executing DAST, and scanning for vulnerabilities. For runtime vulnerability validation. Complements Sentinel static analysis.

Ask your Agent

Use this Skill: https://skilld.dev/gh/simota/agent-skills/probe

This session only. Nothing lands on disk.

referencepentest-methodology-pitfalls.md

≈1.1k tokens on demand. Your agent reads this file only when SKILL.md points to it.

Penetration Testing Methodology Pitfalls

Purpose: Use this file when designing a penetration workflow, validating rules of engagement, or checking whether a test plan is too narrow, too tool-driven, or too weakly evidenced.

Contents

  • PT-01..PT-10
  • Phase-by-phase failure patterns
  • Methodology comparison
  • 2025 attack surfaces
  • Probe quality gates

Top Methodology Pitfalls

ID Anti-pattern Risk Correction
PT-01 Narrow scope Critical assets stay untested Map the full attack surface and prioritize business-critical assets
PT-02 Test without written authorization Legal and ethical exposure Require rules of engagement and approval before testing
PT-03 Skip threat modeling Random testing, weak attack chains Use ATT&CK-informed threat scenarios
PT-04 Over-rely on scanners False positives and missed logic flaws Combine automated scanning with manual validation
PT-05 Stop at detection No proof of exploitability or impact Build a safe detect -> exploit -> impact chain
PT-06 Skip post-exploitation thinking Lateral movement risk stays hidden Assess privilege escalation and impact paths safely
PT-07 Use one framework only Coverage bias Prefer a polymethodology approach
PT-08 Produce a technical-only report Stakeholders cannot act Provide executive and technical layers
PT-09 Skip retest after remediation Fix quality stays unknown Retest confirmed issues after remediation
PT-10 Keep plans static New attack surfaces stay untested Refresh plans for cloud, API, supply chain, and AI changes

Phase Failure Patterns

Phase Common failures
Pre-engagement No written scope, no exclusions, no escalation path
Intelligence gathering No OSINT, no subdomain inventory, no stack fingerprinting
Exploitation No safe proof, no attack chain, weak evidence capture
Reporting No business impact, no remediation, no executive summary

Methodology Comparison

Framework Best use Currency (2026-05)
PTES Lifecycle management and end-to-end pentest structure 2014 baseline still widely cited; supplement with newer guides
OWASP WSTG Web and API technical checks v4.2 (2020-12) remains the current stable; v5.0 in active development at github.com/OWASP/wstg to absorb 2026 AI / WebSocket / serverless realities. No v4.3 has shipped — do not cite it.
MITRE ATT&CK Threat-based scenario design Refresh quarterly from attack.mitre.org
NIST SP 800-115 Compliance-heavy environments 2008 baseline; pair with NIST SP 800-218 SSDF for CI/CD gates
OWASP Top 10 for LLM 2025 v2.0 Generative-AI app surfaces Released 2024-11-18 (genai.owasp.org); use for any LLM-backed endpoint
OWASP Top 10 for Agentic Applications Autonomous AI agents Released 2025-12-09 by OWASP GenAI Security Project; ASI01 Agent Goal Hijacking tops the list

Recommended default: PTES + OWASP WSTG v4.2 + MITRE ATT&CK. Add OWASP Top 10 for LLM 2025 and OWASP Top 10 for Agentic Applications whenever the target embeds LLMs, tool-calling agents, or MCP servers.

2026 Attack Surface Priorities

Area Why it matters (2026-05 evidence)
API ecosystems Wallarm 2026 API ThreatStats: 43% of 2025 CISA KEV additions were API-related (106 / 245); 97% of API vulns exploitable in a single request; BOLA still tops the volume chart
Cloud-native services IAM, serverless, and instance-metadata exposures still rising — pair with nuclei cloud-config templates for GCP / Azure / K8s
Supply chain OWASP A03:2025 Software Supply Chain Failures expanded scope to dependency, build, and CI trust chains; Trivy v0.69.4 incident (2026-03) and dual GitHub Actions compromise illustrate the risk profile
AI / LLM / Agent integrations Wallarm: 2,185 AI-related vulns in 2025, 36% overlap with API vulns; 315 MCP-related vulns with 270% Q2→Q3 growth. Prompt injection (LLM01), indirect injection via RAG, and ASI01 Agent Goal Hijacking now belong in every API-scoped pentest plan

Probe Quality Gates

Condition Required action
No written authorization Block the engagement
No threat model Warn and add scenario design before scanning
Automation only Add manual validation for business logic
No exploit proof Mark as Unconfirmed
No executive summary Report is incomplete

Source: SKILL.md on GitHub

Third-party checks

2 warnings13d5 checks · Risk SAFE
  • Gen Agent Trust Hub13d

    The skill is a comprehensive dynamic security testing specialist designed for vulnerability scanning and penetration testing. No malicious patterns were detected. The flagged prompt injection text consists of industry-standard security test payloads intended for evaluating other systems and does not target the agent's own behavior.

  • Socket13d

    1 alert: gptSecurity

  • Snyk13d

    Risk: LOW · No issues

  • Runlayer6mo

    4/10 files flagged

  • ZeroLeaks5mo

    Score: 93/100 · 2 sections analyzed

Provenance

Signed by skilld at e307415. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub 2 days ago.

Activeupdated 2 weeks ago

README badge

README badge for simota/agent-skills/probe