security-bounty-hunter
Use when the goal is practical vulnerability discovery for responsible disclosure or bounty submission — focuses on remotely reachable, exploitable issues that qualify for real reports rather than a broad best-practices review
Install
npx skills add https://github.com/drvoss/everything-copilot-cli/tree/main/skills/security/security-bounty-hunter
claude plugin marketplace add https://llmmart.ai/marketplace.json && claude plugin install drvoss-everything-copilot-cli@llmmart
git clone https://github.com/drvoss/everything-copilot-cli.git
The skills CLI installs just this skill, for any of its supported agents. Claude Code installs the whole drvoss/everything-copilot-cli collection as a plugin from our marketplace. Git is the plain clone.
Skill manifest
Security Bounty Hunter
Hunt for exploitable, bounty-worthy security issues. Bias toward remotely reachable, user-controlled attack paths — discard patterns that bounty platforms routinely reject as informative or out of scope.
When to Use
- Scanning a repository for exploitable vulnerabilities
- Preparing a Huntr, HackerOne, or similar bounty submission
- Triage where the question is "does this actually qualify for a payout?" rather than "is this theoretically unsafe?"
When NOT to Use
For broad best-practices audits, use security-scan or evaluate-repository instead.
In-Scope Patterns
| Pattern | CWE | Typical impact |
|---|---|---|
| SSRF through user-controlled URLs | CWE-918 | internal network access, cloud metadata theft |
| Auth bypass in middleware or API guards | CWE-287 | unauthorized account or data access |
| Remote deserialization or upload-to-RCE | CWE-502 | code execution |
| SQL injection in reachable endpoints | CWE-89 | data exfiltration, auth bypass |
| Command injection in request handlers | CWE-78 | code execution |
| Path traversal in file-serving paths | CWE-22 | arbitrary file read or write |
| Auto-triggered XSS | CWE-79 | session theft, admin compromise |
Usually Out of Scope (skip unless program says otherwise)
- Local-only
pickle.loads,eval()with no remote path shell=Trueon fully hardcoded commands- Missing security headers alone
- Generic rate-limiting without exploit impact
- Self-XSS requiring victim to paste code manually
- Demo, example, or test-only code
Workflow
- Check scope — program rules, SECURITY.md, disclosure channel, exclusions
- Find real entrypoints — HTTP handlers, uploads, webhooks, background jobs, parsers
- Run static triage — use
grep/powershellfor pattern scanning; treat as triage input only - Read the real code path end to end — follow data from entrypoint to sink
- Prove user control reaches a meaningful sink
- Confirm exploitability with the smallest safe PoC
- Check for duplicates — existing advisories, CVEs, open tickets
Static Triage Patterns
# Find potential injection sinks
grep -r "exec\|eval\|system\|popen\|subprocess" src/ --include="*.py" -n
grep -r "shell=True" src/ --include="*.py" -n
# Find potential SSRF
grep -r "requests.get\|urllib\|fetch\|axios" src/ --include="*.py" --include="*.ts" -n
# Find potential SQL injection
grep -r "f\"SELECT\|f'SELECT\|\+ .* WHERE\|string.*query" src/ -n
Then manually filter: drop tests, demos, fixtures, vendored code, non-reachable paths.
Report Structure
## Description
[What the vulnerability is and why it matters]
## Vulnerable Code
[File path, line range, and a small snippet]
## Proof of Concept
[Minimal working request or script]
## Impact
[What the attacker can achieve]
## Affected Version
[Version, commit, or deployment target tested]
Quality Gate
Before submitting, confirm all of:
- Code path is reachable from a real user or network boundary
- Input is genuinely user-controlled
- Sink is meaningful and exploitable
- PoC works
- Issue is not already covered by an advisory, CVE, or open ticket
- Target is in scope for the bounty program
See Also
- security-scan — OWASP Top 10 + STRIDE broad audit
- evaluate-repository — scored security posture review
- input-validation — sanitize and validate user inputs
Files (everything-copilot-cli)
-
SKILL.md 3.9 KB
--- name: security-bounty-hunter license: MIT description: Use when the goal is practical vulnerability discovery for responsible disclosure or bounty submission — focuses on remotely reachable, exploitable issues that qualify for real reports rather than a broad best-practices review metadata: category: security agent_type: general-purpose origin: ported from affaan-m/everything-claude-code --- # Security Bounty Hunter Hunt for exploitable, bounty-worthy security issues. Bias toward remotely reachable, user-controlled attack paths — discard patterns that bounty platforms routinely reject as informative or out of scope. ## When to Use - Scanning a repository for exploitable vulnerabilities - Preparing a Huntr, HackerOne, or similar bounty submission - Triage where the question is "does this actually qualify for a payout?" rather than "is this theoretically unsafe?" ## When NOT to Use For broad best-practices audits, use [`security-scan`](../security-scan/SKILL.md) or [`evaluate-repository`](../evaluate-repository/SKILL.md) instead. ## In-Scope Patterns | Pattern | CWE | Typical impact | |---------|-----|----------------| | SSRF through user-controlled URLs | CWE-918 | internal network access, cloud metadata theft | | Auth bypass in middleware or API guards | CWE-287 | unauthorized account or data access | | Remote deserialization or upload-to-RCE | CWE-502 | code execution | | SQL injection in reachable endpoints | CWE-89 | data exfiltration, auth bypass | | Command injection in request handlers | CWE-78 | code execution | | Path traversal in file-serving paths | CWE-22 | arbitrary file read or write | | Auto-triggered XSS | CWE-79 | session theft, admin compromise | ## Usually Out of Scope (skip unless program says otherwise) - Local-only `pickle.loads`, `eval()` with no remote path - `shell=True` on fully hardcoded commands - Missing security headers alone - Generic rate-limiting without exploit impact - Self-XSS requiring victim to paste code manually - Demo, example, or test-only code ## Workflow 1. **Check scope** — program rules, SECURITY.md, disclosure channel, exclusions 2. **Find real entrypoints** — HTTP handlers, uploads, webhooks, background jobs, parsers 3. **Run static triage** — use `grep` / `powershell` for pattern scanning; treat as triage input only 4. **Read the real code path end to end** — follow data from entrypoint to sink 5. **Prove user control reaches a meaningful sink** 6. **Confirm exploitability** with the smallest safe PoC 7. **Check for duplicates** — existing advisories, CVEs, open tickets ## Static Triage Patterns ```powershell # Find potential injection sinks grep -r "exec\|eval\|system\|popen\|subprocess" src/ --include="*.py" -n grep -r "shell=True" src/ --include="*.py" -n # Find potential SSRF grep -r "requests.get\|urllib\|fetch\|axios" src/ --include="*.py" --include="*.ts" -n # Find potential SQL injection grep -r "f\"SELECT\|f'SELECT\|\+ .* WHERE\|string.*query" src/ -n ``` Then manually filter: drop tests, demos, fixtures, vendored code, non-reachable paths. ## Report Structure ```markdown ## Description [What the vulnerability is and why it matters] ## Vulnerable Code [File path, line range, and a small snippet] ## Proof of Concept [Minimal working request or script] ## Impact [What the attacker can achieve] ## Affected Version [Version, commit, or deployment target tested] ``` ## Quality Gate Before submitting, confirm all of: - [ ] Code path is reachable from a real user or network boundary - [ ] Input is genuinely user-controlled - [ ] Sink is meaningful and exploitable - [ ] PoC works - [ ] Issue is not already covered by an advisory, CVE, or open ticket - [ ] Target is in scope for the bounty program ## See Also - [security-scan](../security-scan/SKILL.md) — OWASP Top 10 + STRIDE broad audit - [evaluate-repository](../evaluate-repository/SKILL.md) — scored security posture review - [input-validation](../input-validation/SKILL.md) — sanitize and validate user inputs
Comments (0)
Sign in to join the conversation.
Reviews (0)
No reviews yet.
No comments yet.