Install
$ agentstack add skill-mingyiseclab-mingyi-atlas-kill-chain-analysis ✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.
Security review
✓ PassedNo issues found. Passed automated security review. · v0.1.0 How review works →
- ✓ Prompt-injection patterns
- ✓ Secret / credential exfiltration
- ✓ Dangerous shell & filesystem operations
- ✓ Untrusted network calls
- ✓ Known-malicious package signatures
What it can access
- ✓ Network access No
- ✓ Filesystem access No
- ✓ Shell / process execution No
- ✓ Environment & secrets No
- ✓ Dynamic code execution No
From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.
Verified badge
Passed review? Show it. Paste this badge into your README, it links to the public security report.
Reliability & compatibility
Declared compatibility
Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.
We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.
How agent discovery & health will work →About
Kill Chain Analysis & Attack Path Decision-Making
Decision Framework
When selecting the next action, evaluate in order:
- What does the OPPLAN say? — Prioritized objectives drive decisions
- What do findings tell us? — Previous phase results constrain options
- What's the risk/reward? — Lower noise approaches first
- What's the OPSEC impact? — Consult
opsecskill before noisy actions
Findings Analysis
After Recon Phase — Selecting Attack Vectors
Read recon/ outputs and categorize:
| Finding Type | Indicates | Next Action | |-------------|-----------|-------------| | Web apps with known CVEs | Web exploitation path | exploit → web techniques | | AD services (88/389/636) | AD attack surface | exploit → AD techniques (after initial access) | | Exposed credentials (OSINT) | Credential-based access | exploit → credential stuffing/spray | | Cloud misconfigs (S3/blob) | Cloud attack path | exploit → cloud-specific techniques | | VPN/remote access services | Network perimeter entry | exploit → VPN/RDP exploitation | | Employee emails + breach data | Social engineering path | exploit → phishing (if in scope) |
Attack Vector Prioritization
Rank available vectors by:
Score = (Success Probability × Impact) / Detection Risk
1. Valid credentials from OSINT → High prob, High impact, Low noise
2. Known web CVE (public exploit) → High prob, Med impact, Med noise
3. AD misconfiguration (no patch) → Med prob, High impact, Med noise
4. Password spray against O365 → Med prob, High impact, High noise
5. Zero-day or custom exploit → Low prob, High impact, Low noise
Always prefer: credentials > misconfigurations > known CVEs > brute force
After Exploitation — Deciding Post-Exploit Strategy
Once a foothold is established, analyze:
| Context | Decision | |---------|----------| | Low-privilege user on workstation | Prioritize: privesc → cred dump → lateral to server | | Service account on server | Prioritize: cred dump (may have cached admin creds) → lateral | | Domain user credentials | Prioritize: AD enumeration → Kerberoasting → DCSync path | | Local admin on single host | Prioritize: cred dump → check for cached domain creds → lateral | | Already domain admin | Prioritize: objective completion → evidence collection → reporting |
Handling Blocked Objectives
Failure Analysis Decision Tree
Objective BLOCKED
│
├── WHY did it fail?
│ ├── Defense mechanism (WAF/EDR/IDS)
│ │ → Consult defense-evasion skill → retry with evasion
│ │
│ ├── Missing prerequisite (need creds/access/info)
│ │ → Identify which prior phase provides it → re-order objectives
│ │
│ ├── Target hardened / not vulnerable
│ │ → Check findings for alternative target → redirect attack
│ │
│ └── Tool failure / environment issue
│ → Retry with different tool or approach
│
├── Is there an ALTERNATIVE path?
│ ├── YES → Craft new delegation with adjusted approach
│ └── NO → Mark BLOCKED, document reason, proceed to next objective
│
└── Should we REVISIT later?
├── YES (new intel may help) → Keep status BLOCKED, note in lessons_learned.md
└── NO (dead end) → Mark BLOCKED permanently
Common Pivots
| Original Approach | Alternative When Blocked | |-------------------|-------------------------| | SQLi on web app | SSTI, deserialization, SSRF, or move to different web app | | Kerberoasting | AS-REP roasting, ADCS abuse, password spray | | LSASS dump blocked by EDR | nanodump, comsvcs.dll, MiniDumpWriteDump via syscall | | WinRM blocked | PsExec, WMI, DCOM, RDP, SMB exec | | Password spray lockout | Low-and-slow spray, single-password-multiple-users |
MITRE ATT&CK Phase Mapping
Use this to map OPPLAN objective phases to ATT&CK tactics:
| OPPLAN Phase | ATT&CK Tactic | Sub-Agent | |-------------|---------------|-----------| | Recon | TA0043 Reconnaissance | recon | | Initial Access | TA0001 Initial Access | exploit | | Execution | TA0002 Execution | exploit | | Persistence | TA0003 Persistence | postexploit | | Privilege Escalation | TA0004 Privilege Escalation | postexploit | | Defense Evasion | TA0005 Defense Evasion | exploit / postexploit | | Credential Access | TA0006 Credential Access | postexploit | | Discovery | TA0007 Discovery | recon / postexploit | | Lateral Movement | TA0008 Lateral Movement | postexploit | | Collection | TA0009 Collection | postexploit | | Exfiltration | TA0010 Exfiltration | postexploit |
Target Prioritization
When multiple targets are available, prioritize:
- Crown jewels — Targets explicitly named in OPPLAN objectives
- Domain controllers — Access = domain-wide compromise
- File servers — Likely contain sensitive data for objectives
- Admin workstations — Cached credentials for further access
- Application servers — May contain database credentials, API keys
- Standard workstations — Lower value, but may provide stepping stones
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: MingyiSecLab
- Source: MingyiSecLab/Mingyi-Atlas
- License: Apache-2.0
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet, be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.