vulnerability-chaining
Skill47 starsChanged 5 months agoMITView on GitHub
Use when you have multiple findings and need to combine them into realistic end-to-end attack paths — turning several low/medium issues into a single critical-impact exploit chain. The red-team-lead skill that ties the specialists' findings together into business impact.
In this repository15 other files it gives its agents.
- access-control-testingSkill
- active-pentestSkill
- api-security-testingSkill
- authentication-testingSkill
- business-logic-testingSkill
- client-side-exploitationSkill
- file-upload-and-ssrfSkill
- injection-testingSkill
- pentest-reportingSkill
- recon-and-osintSkill
- secrets-management-auditSkill
- security-code-auditSkill
- security-hardeningSkill
- security-testingSkill
- threat-modelingSkill
The file
Files
Vulnerability Chaining (the team lead)
Overview
Individual findings undersell real risk. Attackers don't stop at one bug — they chain them: an info leak feeds an enumeration, which feeds an auth bypass, which reaches admin. This skill takes the register from all FIND/EXPLOIT specialists and assembles attack paths.
Core principle: Chain impact is multiplicative, not additive. Three "Mediums" that connect can equal one "Critical." Severity belongs to the path, not the isolated bug.
⚠️ Authorization
Demonstrate chains only on your own/authorized non-prod instance. Prove the path with the minimum steps; stop at proof of impact — no persistence, no real-data exfiltration, no pivoting to systems out of scope.
Method
- Inventory primitives — re-read every finding as a capability, not a label:
- Info (enumeration, verbose errors, leaked ids/secrets, source maps)
- Foothold (any auth bypass, XSS, SSRF, upload)
- Escalation (IDOR→admin, privesc, mass assignment)
- Pivot (SSRF→metadata, leaked key→cloud, token reuse)
- Find adjacencies — output of one bug = precondition of another? (leaked email → no-rate-limit login → ATO; reflected XSS → steals admin session → admin IDOR).
- Build the path — order primitives start→goal (unauth → user → admin → data/RCE).
- Demonstrate safely — walk the chain end-to-end on staging, capturing each hop's PoC.
- Score the chain — rate by the final impact, then list the constituent bugs.
See references/attack-paths.md for canonical chain templates.
Quick reference — high-value chains
| Chain | Path |
| Full account takeover | user enumeration → no rate limit → weak reset/JWT → ATO |
| XSS → admin compromise | stored XSS → steal admin cookie → admin BFLA → mass data |
| SSRF → cloud takeover | SSRF → 169.254.169.254 metadata → IAM creds → cloud API |
| IDOR → privesc | IDOR leaks admin id/email → password reset on admin → admin access |
| Upload → RCE | unrestricted upload → web-accessible path → execute → shell |
| Logic → fraud | coupon race → negative balance → withdraw → financial loss |
Output
One narrative per chain: the attack path (numbered hops), the constituent finding IDs, the end impact in business terms, and the single point where breaking the chain stops it (the cheapest fix). This is what makes leadership act.
Hand-off
Findings come from all specialists; chains go to pentest-reporting as your top criticals,
and to security-hardening highlighting the cheapest chain-breaking fix.
Common mistakes
- Reporting bugs in isolation — you'll under-rate the real risk and the chain gets ignored.
- Building theoretical chains — each hop must be demonstrated, not assumed.
- Over-running the chain — stop at proven impact; don't persist or pivot out of scope.
- Missing the cheapest break — often one mid-chain fix kills several paths at once.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.