| name | your-skill-name |
| description | Use when [trigger conditions]. Performs [what it does] for [target language/framework]. |
Your Skill Name
One paragraph explaining the skill's purpose. For example: "This skill performs static analysis
for [vulnerability class] in [language/framework] projects, identifying [specific issues] and
mapping findings to CWE/OWASP standards."
When to Use
- When the user asks to...
- When scanning code for...
- When reviewing...
- When a pull request contains changes to...
When NOT to Use
- When the user is asking about...
- For non-security-related...
- When another skill (e.g.,
skill-name) already covers...
Prerequisites
Tool Installed (Preferred)
which toolname || toolname --version
npm install -g toolname
Tool Not Installed (Fallback)
Note: Manual checks are less comprehensive than automated tooling. Consider installing
toolname for full coverage.
- Check for [common vulnerability pattern 1]
- Check for [common vulnerability pattern 2]
- Check for [common vulnerability pattern 3]
- Check for [common vulnerability pattern 4]
- Check for [common vulnerability pattern 5]
- Check for [common vulnerability pattern 6]
- Check for [common vulnerability pattern 7]
- Check for [common vulnerability pattern 8]
- Check for [common vulnerability pattern 9]
- Check for [common vulnerability pattern 10]
Workflow
- Detect project language/framework by inspecting package.json, requirements.txt, go.mod, etc.
- Identify target files for scanning (e.g.,
src/**/*.ts, app/**/*.py)
- Scan for [specific vulnerability patterns]
- For each finding:
a. Determine severity (Critical / High / Medium / Low)
b. Map to the relevant CWE identifier
c. Map to the relevant OWASP Top 10 category
d. Identify the file and line number
e. Draft a remediation recommendation
- Deduplicate and sort findings by severity (Critical first)
- Generate the findings report using the format below
- Summarize: total findings, breakdown by severity, top recommendations
Findings Format
Each finding should include:
| Field | Description |
|---|
| Severity | Critical / High / Medium / Low |
| CWE | CWE-XXX identifier |
| OWASP | A01-A10 category |
| Location | file:line |
| Issue | Description of the vulnerability |
| Remediation | How to fix it |
Example Finding
| Field | Value |
|---|
| Severity | High |
| CWE | CWE-79 |
| OWASP | A03:2021 - Injection |
| Location | src/components/UserProfile.tsx:42 |
| Issue | User-supplied name is rendered without sanitization via dangerouslySetInnerHTML |
| Remediation | Use React's default text rendering or sanitize input with DOMPurify before rendering |
Reference Tables
| Check | CWE | OWASP | Severity |
|---|
| Example check 1 | CWE-79 | A03 | High |
| Example check 2 | CWE-89 | A03 | Critical |
| Example check 3 | CWE-200 | A01 | Medium |
| Example check 4 | CWE-522 | A07 | High |
Example Usage
User prompt:
"Run a [your-skill-name] scan on this project"
Expected output (abbreviated):
## [Your Skill Name] Scan Results
Scanned 23 files in src/
### Findings (3 total: 1 Critical, 1 High, 1 Medium)
| # | Severity | CWE | OWASP | Location | Issue |
|---|----------|-----|-------|----------|-------|
| 1 | Critical | CWE-89 | A03 | src/db/queries.ts:18 | SQL query built via string concatenation with user input |
| 2 | High | CWE-79 | A03 | src/views/profile.ejs:7 | Unescaped user output in template |
| 3 | Medium | CWE-200 | A01 | src/errors/handler.ts:34 | Stack trace exposed in production error response |
### Recommendations
1. Use parameterized queries for all database access (Finding #1)
2. Enable auto-escaping in EJS templates (Finding #2)
3. Suppress stack traces when NODE_ENV=production (Finding #3)
