Common Dupes Database

Usage

/greyhatcc:dupes <finding_description> /greyhatcc:dupes list — show all patterns

Checks a finding description against a curated database of findings that are commonly rejected across bug bounty programs. Saves you from writing a report that will be marked N/A, Informational, or Won't Fix.

Context Loading (MANDATORY)

Before executing this skill:

1. Load scope: .greyhatcc/scope.json — verify target is in scope, note exclusions
2. Load hunt state: .greyhatcc/hunt-state.json — check active phase, resume context
3. Load program files: findings_log.md, tested.json, gadgets.json — avoid duplicating work
4. Load memory: Check MEMORY.md for target-specific notes from previous sessions

How It Works

The database is implemented in scripts/lib/dupes.mjs and contains 24+ patterns categorized by rejection confidence:

Confidence Levels

Level	Rejection Rate	What To Do
ALWAYS_REJECTED	95%+	Don't submit. Only worth it if chained into something bigger.
USUALLY_REJECTED	70-95%	Need very strong PoC + demonstrated impact to overcome.
SOMETIMES_REJECTED	40-70%	Program-dependent. Check program's exclusion list and hacktivity.
CONTEXT_DEPENDENT	Varies	Entirely depends on the depth of your proof and impact.

ALWAYS_REJECTED (Don't Even Bother)

Pattern	Why Rejected	Chain Opportunity
Missing HSTS	Universally excluded	HTTP downgrade + session hijack + cookie theft
Missing X-Frame-Options / Clickjacking	Universally excluded	Clickjack + sensitive action (password change)
Missing/weak CSP	Not a vulnerability alone	XSS + CSP bypass = full exploitation
Missing SPF/DKIM/DMARC	Not a web vulnerability	None — move on
Missing cookie flags	Universally excluded	XSS + missing HttpOnly = session theft
robots.txt / sitemap.xml disclosure	Public by design	None — move on
Weak SSL/TLS ciphers	Universally excluded	None practical
Server version disclosure	Not a vulnerability alone	Version + known CVE with working exploit
Self-XSS	Requires victim action	Self-XSS + CSRF = forced execution
Root/jailbreak detection bypass	The bypass isn't the bug	What you find AFTER bypass is the bug

USUALLY_REJECTED (Need Strong Proof)

Pattern	Why Usually Rejected	How to Overcome
Open redirect (standalone)	Low impact alone	Chain with OAuth token theft or SSRF
CORS without data exfil PoC	Theoretical without proof	Build working HTML page that reads data cross-origin
User enumeration	"By design" for most programs	Show bulk enumeration of ALL users or admin accounts
Content/HTML injection	No script execution	Escalate to stored XSS or credential phishing
CSRF on login/logout/preferences	Low impact	Chain login CSRF + self-XSS
Missing rate limiting (generic)	Excluded unless exploitable	Show OTP brute force, account takeover, or financial drain
Verbose errors / stack traces	Informational	Chain with leaked internal paths + LFI/SSRF
Outdated library without exploit	Theoretical	Write a working exploit against the target

SOMETIMES_REJECTED (Program-Dependent)

Pattern	When Accepted	When Rejected
Host header injection	Cache poisoning or password reset hijack	Simple reflection without impact
Subdomain takeover	Demonstrated claim with proof page	Dangling CNAME without claim proof
GraphQL introspection	Reveals sensitive mutations/data	Schema is already documented
Exposed actuator	/env, /heapdump leak secrets	/health and /info only

CONTEXT_DEPENDENT (Quality of Proof Decides)

Pattern	Critical Version	N/A Version
IDOR	Cross-user PII access with enumerable IDs	Reading your own data with different ID format
SSRF	Cloud metadata + IAM cred extraction	DNS callback only, no response data
CSRF	Account takeover via password/email change	Preference toggle, language change
Prototype pollution	Server-side PP → RCE, or client-side PP → XSS	PP in library with no exploitable gadget
Information disclosure	Leaked credentials, PII, internal IPs leading to SSRF	Generic error messages, public repo info
Race condition	Double-spend on payments, OTP brute force	Like button count inflation, non-critical counters

Program-Specific Exceptions

Some programs have unique acceptance/rejection patterns. Always check the specific program's exclusion list and hacktivity:

Pattern	Exception Programs	Notes
Open redirect	Some fintech programs accept standalone	Financial apps: phishing risk is higher
User enumeration	Healthcare/HIPAA-regulated programs	Patient/user existence is sensitive in healthcare
Missing rate limiting	Programs with payment/financial flows	Brute force on OTP = account takeover
GraphQL introspection	Programs with sensitive internal schemas	When schema reveals admin mutations or PII fields
Subdomain takeover (no claim)	Some programs accept evidence-only	If program policy says "do not claim subdomains"

Chain-Only Escalation Paths

When a finding matches a rejected pattern, these are the escalation paths to turn it into an accepted report:

Rejected Pattern	Chain With	Escalated To
Self-XSS	CSRF on same form	Forced XSS execution → session hijack (HIGH)
Open redirect	OAuth redirect_uri on same domain	Token theft → ATO (CRITICAL)
Missing cookie flags (HttpOnly)	XSS on same domain	Session cookie theft via document.cookie (HIGH)
Missing X-Frame-Options	Sensitive action page (password change, transfer)	Clickjacking → unauthorized action (MEDIUM-HIGH)
User enumeration	Credential stuffing with breach data	Bulk account compromise (HIGH)
CORS without exfil	Subdomain takeover on trusted origin	Cross-origin authenticated data theft (HIGH)
Content injection	Credential phishing via injected login form	Credential theft (MEDIUM-HIGH)
Server version disclosure	Known CVE with working exploit for that version	RCE/auth bypass (CRITICAL)
Missing rate limiting	OTP/2FA endpoint brute force	Account takeover via OTP bypass (HIGH)
Verbose errors/stack traces	Leaked internal paths + LFI/SSRF on same host	Internal file read / SSRF pivot (HIGH)
GraphQL introspection	Hidden admin mutations + auth bypass	Unauthorized admin actions (HIGH-CRITICAL)
Root/jailbreak detection bypass	Discovered hardcoded credentials in app	Credential theft → backend access (HIGH)
SSL pinning bypass	MitM reveals undocumented API with auth flaws	API exploitation (varies)

Usage in Code

The database is also available programmatically via the finding-tracker hook and report-validator hook. Both automatically check findings against these patterns and warn before you invest time in a report.

Maintenance

This database should be updated when:

• A new finding type gets consistently rejected across multiple programs
• A previously-rejected type starts getting accepted (new attack chains, new impact demonstrations)
• Program-specific exclusion patterns are identified

Delegation

• No delegation needed — this is a lookup tool
• Results feed into: /greyhatcc:dedup, /greyhatcc:validate, report-validator hook

Agent Dispatch Protocol

When delegating to agents via Task(), ALWAYS:

1. Prepend worker preamble: "[WORKER] Execute directly. No sub-agents. Output ≤500 words. Save findings to disk. 3 failures = stop and report."
2. Set max_turns: haiku=10, sonnet=25, opus=40
3. Pass full context: scope, exclusions, existing findings, recon data
4. Route by complexity: Quick checks → haiku agents (-low). Standard work → sonnet agents. Deep analysis/exploitation → opus agents.

State Updates

After completing this skill:

1. Update tested.json — record what was tested (asset + vuln class)
2. Update gadgets.json — add any informational findings with provides/requires tags for chaining
3. Update findings_log.md — log any confirmed findings with severity
4. Update hunt-state.json if in active hunt — set lastActivity timestamp