You are in AUTONOMOUS MODE. Do NOT ask questions. Hunt for exploitable vulnerabilities methodically.
TARGET:
$ARGUMENTS
If no arguments provided, perform a full code-level penetration test of the project in the current working directory. If arguments specify a scope (e.g., "auth module", "API routes"), focus the analysis on that area.
============================================================
PHASE 0: RECONNAISSANCE
Map the application's attack surface:
TECH STACK:
- Detect framework, language, database, auth mechanism
- Identify entry points (HTTP routes, WebSocket handlers, CLI commands, queue consumers)
- Map middleware chain (auth, validation, rate limiting, logging)
ROUTE INVENTORY:
- List ALL defined routes/endpoints with their HTTP methods
- Note which routes have authentication middleware
- Note which routes accept user input (body, query, params, headers, files)
- Flag routes that perform privileged operations
DATA FLOW MAPPING:
- Trace user input from entry point through processing to output/storage
- Identify where input is validated, sanitized, or escaped
- Identify where input reaches dangerous sinks (database, shell, filesystem, HTML)
Record the full attack surface map before beginning targeted analysis.
============================================================
PHASE 1: SQL / NOSQL INJECTION
Hunt for injection points in database interactions:
SQL INJECTION:
- Search for string concatenation in SQL:
"SELECT * FROM " + table
- Search for template literals in SQL:
`SELECT * FROM ${table}`
- Search for format strings in SQL:
"SELECT * FROM %s" % table
- Check ORM raw query methods for unsanitized input
- Check stored procedures called with dynamic parameters
- Verify parameterized queries are used consistently
NOSQL INJECTION:
- MongoDB: Search for
$where, $regex, $gt, $ne with user input
- MongoDB: Check for object injection via
JSON.parse(userInput) in queries
- Firestore: Check for user-controlled field paths in queries
- Redis: Check for user input in command strings
For each finding:
- Location: file:line
- Severity: Critical
- Proof-of-concept: Example malicious input that would exploit the flaw
- Fix: Parameterized query replacement code
============================================================
PHASE 2: CROSS-SITE SCRIPTING (XSS)
Hunt for XSS vectors:
REFLECTED XSS:
- User input reflected in HTML responses without encoding
- Search parameters displayed on results pages
- Error messages containing user input
- URL path segments rendered in page content
STORED XSS:
- User-generated content (comments, profiles, messages) rendered without escaping
- File names displayed without sanitization
- User-uploaded HTML/SVG files served inline
DOM-BASED XSS:
document.write() with URL fragments or query parameters
innerHTML assignment with user-controlled data
eval() or Function() with user input
location.href / window.open() with user-controlled URLs
- jQuery
.html() with unescaped data
TEMPLATE INJECTION:
- Server-side template injection (SSTI): user input in template strings
- Client-side:
dangerouslySetInnerHTML, v-html, [innerHTML], {!! !!}
- Missing auto-escape:
<%- %> (EJS), | safe (Jinja2), {{{ }}} (Handlebars)
For each finding:
- Location: file:line
- Severity: High-Critical
- Proof-of-concept:
<script>alert(1)</script> or context-specific payload
- Fix: Proper encoding/escaping code
============================================================
PHASE 3: AUTHENTICATION BYPASS
Hunt for authentication weaknesses:
MISSING AUTH:
- Routes handling sensitive data without auth middleware
- Admin endpoints accessible without admin role check
- API endpoints that should require auth but do not
- WebSocket connections without authentication handshake
AUTH LOGIC FLAWS:
- JWT verification that can be bypassed (algorithm confusion, missing expiry check)
- Session tokens predictable or not cryptographically random
- Password reset tokens reusable or long-lived
- OAuth state parameter missing (CSRF on OAuth flow)
- Remember-me tokens without proper validation
CREDENTIAL ATTACKS:
- No rate limiting on login endpoint
- No account lockout after failed attempts
- User enumeration via different error messages (valid vs invalid email)
- Timing attacks on authentication (measurably different response times)
For each finding:
- Location: file:line or route definition
- Severity: Critical-High
- Proof-of-concept: Steps to bypass authentication
- Fix: Middleware addition or logic correction
============================================================
PHASE 4: AUTHORIZATION FLAWS
Hunt for privilege escalation:
HORIZONTAL PRIVILEGE ESCALATION:
- Accessing other users' resources by changing ID parameters
- Missing ownership checks:
findById(req.params.id) without where: { userId: req.user.id }
- Batch endpoints that do not filter by the requesting user
- File access without ownership verification
VERTICAL PRIVILEGE ESCALATION:
- Admin functions accessible by regular users
- Role changes possible without admin authorization
- Missing role checks on mutation endpoints (only checked on read)
- Client-side-only role enforcement (no server-side check)
MASS ASSIGNMENT:
- Request body spread directly into database updates:
Model.update(req.body)
- User can set
role, isAdmin, balance via API request
- GraphQL mutations accepting arbitrary fields
For each finding:
- Location: file:line
- Severity: Critical-High
- Proof-of-concept: Example API call exploiting the flaw
- Fix: Authorization middleware or ownership check code
============================================================
PHASE 5: PATH TRAVERSAL AND FILE INCLUSION
Hunt for file system attacks:
PATH TRAVERSAL:
- User input in
fs.readFile(), open(), File() without path sanitization
- File download endpoints using user-supplied filenames
- Pattern:
path.join(baseDir, userInput) without ../ prevention
- Image/document serving with user-controlled paths
FILE UPLOAD:
- Missing file type validation (only checking extension, not magic bytes)
- Missing file size limits
- Uploaded files stored in web-accessible directories
- Uploaded filenames not sanitized (may contain
../ or special chars)
- No virus/malware scanning on uploads
LOCAL FILE INCLUSION:
- Dynamic
require() or import() with user input
- Template file paths from user input
- Configuration file paths from user input
For each finding:
- Location: file:line
- Severity: High-Critical
- Proof-of-concept:
../../etc/passwd or equivalent payload
- Fix: Path sanitization code
============================================================
PHASE 6: COMMAND INJECTION
Hunt for command injection:
DIRECT INJECTION:
child_process.exec() with user input: exec("convert " + filename)
os.system(), subprocess.run(shell=True) with user arguments
Runtime.exec() with unsanitized parameters
- Backtick execution with user data
INDIRECT INJECTION:
- User input in environment variables passed to subprocesses
- User-controlled arguments in CLI tool invocations
- CRON job expressions from user input
- Git commands with user-supplied branch/repo names
For each finding:
- Location: file:line
- Severity: Critical
- Proof-of-concept:
; cat /etc/passwd or $(whoami) payload
- Fix: Use array-form exec, escape arguments, or avoid shell execution
============================================================
PHASE 7: CSRF AND REQUEST FORGERY
Hunt for CSRF vulnerabilities:
MISSING CSRF PROTECTION:
- State-changing POST/PUT/DELETE endpoints without CSRF tokens
- Missing
SameSite cookie attribute
- CORS allowing credentials with broad origins
- Form submissions without anti-CSRF tokens
SSRF:
- User-supplied URLs passed to HTTP clients (fetch, axios, requests)
- URL redirects without origin validation
- Webhook URLs from user input fetched server-side
- Image/preview URLs fetched without allowlist
OPEN REDIRECTS:
- Redirect URLs from query parameters without validation
- Login redirect (
?next= or ?redirect=) without allowlist
- OAuth callback URLs not properly validated
For each finding:
- Location: file:line
- Severity: Medium-High
- Proof-of-concept: Example malicious URL or forged request
- Fix: CSRF token implementation, URL validation code
============================================================
PHASE 8: SECRETS AND CREDENTIALS
Hunt for hardcoded secrets:
PATTERNS TO SEARCH:
- API keys: strings matching
[A-Za-z0-9_-]{20,} near key-related variables
- AWS keys:
AKIA[0-9A-Z]{16}
- Private keys:
-----BEGIN (RSA |EC |)PRIVATE KEY-----
- JWT secrets: hardcoded strings in JWT sign/verify calls
- Database passwords: connection strings with embedded credentials
- OAuth client secrets: hardcoded in source (not env vars)
- Encryption keys: hardcoded byte arrays or strings used as keys
CHECK CONFIGURATION:
.env files committed (not in .gitignore)
- Secrets in
docker-compose.yml (not using Docker secrets)
- Secrets in CI/CD config files (not using secret variables)
- Secrets in client-side/frontend code (visible to users)
For each finding:
- Location: file:line
- Severity: Critical
- Secret type: [what kind of credential]
- Fix: Move to environment variable or secret manager
============================================================
PHASE 9: INSECURE DESERIALIZATION
Hunt for deserialization attacks:
JSON.parse() on untrusted input used to construct queries or commands
pickle.loads() on user-supplied data (Python)
yaml.load() without SafeLoader (Python)
unserialize() on user input (PHP)
ObjectInputStream with untrusted data (Java)
Marshal.load() on untrusted data (Ruby)
- XML parsing without disabling external entities (XXE)
For each finding:
- Location: file:line
- Severity: High-Critical
- Proof-of-concept: Malicious serialized payload structure
- Fix: Safe deserialization method or input validation
============================================================
SELF-HEALING VALIDATION (max 2 iterations)
After producing the security analysis, validate thoroughness:
- Verify every category in the audit was actually checked (not skipped).
- Verify every finding has a specific file:line location.
- Verify severity ratings are justified by impact assessment.
- Verify no false positives by re-reading flagged code in context.
IF VALIDATION FAILS:
- Re-audit skipped categories or vague findings
- Verify or remove false positives
- Repeat up to 2 iterations
============================================================
OUTPUT
Penetration Test Report (Code Analysis)
Project: [name]
Stack: [detected technologies]
Scope: [full project or specified scope]
Date: [date]
Executive Summary
| Category |
Critical |
High |
Medium |
Low |
| Injection (SQL/NoSQL) |
N |
N |
N |
N |
| XSS |
N |
N |
N |
N |
| Auth Bypass |
N |
N |
N |
N |
| Authorization Flaws |
N |
N |
N |
N |
| Path Traversal |
N |
N |
N |
N |
| Command Injection |
N |
N |
N |
N |
| CSRF/SSRF |
N |
N |
N |
N |
| Hardcoded Secrets |
N |
N |
N |
N |
| Insecure Deserialization |
N |
N |
N |
N |
| Total |
N |
N |
N |
N |
Findings
For each finding, in severity order:
[SEVERITY] — [Finding Title]
Attack Surface Summary
- Total routes: N
- Unauthenticated routes: N
- Routes accepting user input: N
- Routes with dangerous sinks: N
Remediation Priority
[Ordered list: Critical findings first, with effort estimates]
============================================================
NEXT STEPS
After reviewing the pentest report:
- "Fix all Critical findings immediately — these are exploitable."
- "Run
/owasp for a systematic OWASP Top 10 compliance check."
- "Run
/encryption to address cryptographic weaknesses."
- "Run
/dependency-scan to fix vulnerable component findings."
- "Run
/secure after remediation to verify the security posture improved."
============================================================
SELF-EVOLUTION TELEMETRY
After producing output, record execution metadata for the /evolve pipeline.
Check if a project memory directory exists:
- Look for the project path in
~/.claude/projects/
- If found, append to
skill-telemetry.md in that memory directory
Entry format:
### /pentest — {{YYYY-MM-DD}}
- Outcome: {{SUCCESS | PARTIAL | FAILED}}
- Self-healed: {{yes — what was healed | no}}
- Iterations used: {{N}} / {{N max}}
- Bottleneck: {{phase that struggled or "none"}}
- Suggestion: {{one-line improvement idea for /evolve, or "none"}}
Only log if the memory directory exists. Skip silently if not found.
Keep entries concise — /evolve will parse these for skill improvement signals.
============================================================
DO NOT
- Do NOT execute actual exploits — this is static code analysis only.
- Do NOT send network requests to test vulnerabilities.
- Do NOT expose full secret values — redact all but first 4 characters.
- Do NOT modify any code — report findings with suggested fixes only.
- Do NOT report theoretical vulnerabilities without code evidence — every finding must cite a specific file and line.
- Do NOT conflate "possible" with "confirmed" — clearly state confidence level.
- Do NOT skip any phase — even if the project seems small, check all categories.
- Do NOT install external tools — analyze the source code directly.
1---2name: pentest3description: "Static-analysis penetration test that hunts for exploitable vulnerabilities with proof-of-concept payloads and fix code. Covers SQL and NoSQL injection (string concatenation, raw queries, operator injection), XSS (reflected, stored, DOM-based, template injection, dangerouslySetInnerHTML), authentication bypass (missing auth middleware, JWT algorithm confusion, predictable tokens, OAuth state CSRF), authorization flaws (IDOR, mass assignment, horizontal/vertical privilege escalation), path traversal and file inclusion (unsanitized file paths, upload validation, LFI), command injection (exec, system, subprocess with user input), CSRF and SSRF (missing SameSite, user-supplied URLs, open redirects), hardcoded secrets (AWS keys, private keys, JWT secrets, connection strings, .env in git), and insecure deserialization (pickle, yaml.load, XXE, ObjectInputStream). Maps full attack surface with route inventory. Use for pre-release security validation, finding exploitable bugs, or generating penetration test evidence."4version: "2.0.0"5category: security6platforms:7 - CLAUDE_CODE8metadata:9 source: https://skillmd.com/skills/tinh2/pentest10---1112You are in AUTONOMOUS MODE. Do NOT ask questions. Hunt for exploitable vulnerabilities methodically.1314TARGET:15$ARGUMENTS1617If no arguments provided, perform a full code-level penetration test of the project in the current working directory. If arguments specify a scope (e.g., "auth module", "API routes"), focus the analysis on that area.1819============================================================20PHASE 0: RECONNAISSANCE21============================================================2223Map the application's attack surface:2425TECH STACK:26- Detect framework, language, database, auth mechanism27- Identify entry points (HTTP routes, WebSocket handlers, CLI commands, queue consumers)28- Map middleware chain (auth, validation, rate limiting, logging)2930ROUTE INVENTORY:31- List ALL defined routes/endpoints with their HTTP methods32- Note which routes have authentication middleware33- Note which routes accept user input (body, query, params, headers, files)34- Flag routes that perform privileged operations3536DATA FLOW MAPPING:37- Trace user input from entry point through processing to output/storage38- Identify where input is validated, sanitized, or escaped39- Identify where input reaches dangerous sinks (database, shell, filesystem, HTML)4041Record the full attack surface map before beginning targeted analysis.4243============================================================44PHASE 1: SQL / NOSQL INJECTION45============================================================4647Hunt for injection points in database interactions:4849SQL INJECTION:50- Search for string concatenation in SQL: `"SELECT * FROM " + table`51- Search for template literals in SQL: `` `SELECT * FROM ${table}` ``52- Search for format strings in SQL: `"SELECT * FROM %s" % table`53- Check ORM raw query methods for unsanitized input54- Check stored procedures called with dynamic parameters55- Verify parameterized queries are used consistently5657NOSQL INJECTION:58- MongoDB: Search for `$where`, `$regex`, `$gt`, `$ne` with user input59- MongoDB: Check for object injection via `JSON.parse(userInput)` in queries60- Firestore: Check for user-controlled field paths in queries61- Redis: Check for user input in command strings6263For each finding:64- **Location:** file:line65- **Severity:** Critical66- **Proof-of-concept:** Example malicious input that would exploit the flaw67- **Fix:** Parameterized query replacement code6869============================================================70PHASE 2: CROSS-SITE SCRIPTING (XSS)71============================================================7273Hunt for XSS vectors:7475REFLECTED XSS:76- User input reflected in HTML responses without encoding77- Search parameters displayed on results pages78- Error messages containing user input79- URL path segments rendered in page content8081STORED XSS:82- User-generated content (comments, profiles, messages) rendered without escaping83- File names displayed without sanitization84- User-uploaded HTML/SVG files served inline8586DOM-BASED XSS:87- `document.write()` with URL fragments or query parameters88- `innerHTML` assignment with user-controlled data89- `eval()` or `Function()` with user input90- `location.href` / `window.open()` with user-controlled URLs91- jQuery `.html()` with unescaped data9293TEMPLATE INJECTION:94- Server-side template injection (SSTI): user input in template strings95- Client-side: `dangerouslySetInnerHTML`, `v-html`, `[innerHTML]`, `{!! !!}`96- Missing auto-escape: `<%- %>` (EJS), `| safe` (Jinja2), `{{{ }}}` (Handlebars)9798For each finding:99- **Location:** file:line100- **Severity:** High-Critical101- **Proof-of-concept:** `<script>alert(1)</script>` or context-specific payload102- **Fix:** Proper encoding/escaping code103104============================================================105PHASE 3: AUTHENTICATION BYPASS106============================================================107108Hunt for authentication weaknesses:109110MISSING AUTH:111- Routes handling sensitive data without auth middleware112- Admin endpoints accessible without admin role check113- API endpoints that should require auth but do not114- WebSocket connections without authentication handshake115116AUTH LOGIC FLAWS:117- JWT verification that can be bypassed (algorithm confusion, missing expiry check)118- Session tokens predictable or not cryptographically random119- Password reset tokens reusable or long-lived120- OAuth state parameter missing (CSRF on OAuth flow)121- Remember-me tokens without proper validation122123CREDENTIAL ATTACKS:124- No rate limiting on login endpoint125- No account lockout after failed attempts126- User enumeration via different error messages (valid vs invalid email)127- Timing attacks on authentication (measurably different response times)128129For each finding:130- **Location:** file:line or route definition131- **Severity:** Critical-High132- **Proof-of-concept:** Steps to bypass authentication133- **Fix:** Middleware addition or logic correction134135============================================================136PHASE 4: AUTHORIZATION FLAWS137============================================================138139Hunt for privilege escalation:140141HORIZONTAL PRIVILEGE ESCALATION:142- Accessing other users' resources by changing ID parameters143- Missing ownership checks: `findById(req.params.id)` without `where: { userId: req.user.id }`144- Batch endpoints that do not filter by the requesting user145- File access without ownership verification146147VERTICAL PRIVILEGE ESCALATION:148- Admin functions accessible by regular users149- Role changes possible without admin authorization150- Missing role checks on mutation endpoints (only checked on read)151- Client-side-only role enforcement (no server-side check)152153MASS ASSIGNMENT:154- Request body spread directly into database updates: `Model.update(req.body)`155- User can set `role`, `isAdmin`, `balance` via API request156- GraphQL mutations accepting arbitrary fields157158For each finding:159- **Location:** file:line160- **Severity:** Critical-High161- **Proof-of-concept:** Example API call exploiting the flaw162- **Fix:** Authorization middleware or ownership check code163164============================================================165PHASE 5: PATH TRAVERSAL AND FILE INCLUSION166============================================================167168Hunt for file system attacks:169170PATH TRAVERSAL:171- User input in `fs.readFile()`, `open()`, `File()` without path sanitization172- File download endpoints using user-supplied filenames173- Pattern: `path.join(baseDir, userInput)` without `../` prevention174- Image/document serving with user-controlled paths175176FILE UPLOAD:177- Missing file type validation (only checking extension, not magic bytes)178- Missing file size limits179- Uploaded files stored in web-accessible directories180- Uploaded filenames not sanitized (may contain `../` or special chars)181- No virus/malware scanning on uploads182183LOCAL FILE INCLUSION:184- Dynamic `require()` or `import()` with user input185- Template file paths from user input186- Configuration file paths from user input187188For each finding:189- **Location:** file:line190- **Severity:** High-Critical191- **Proof-of-concept:** `../../etc/passwd` or equivalent payload192- **Fix:** Path sanitization code193194============================================================195PHASE 6: COMMAND INJECTION196============================================================197198Hunt for command injection:199200DIRECT INJECTION:201- `child_process.exec()` with user input: `exec("convert " + filename)`202- `os.system()`, `subprocess.run(shell=True)` with user arguments203- `Runtime.exec()` with unsanitized parameters204- Backtick execution with user data205206INDIRECT INJECTION:207- User input in environment variables passed to subprocesses208- User-controlled arguments in CLI tool invocations209- CRON job expressions from user input210- Git commands with user-supplied branch/repo names211212For each finding:213- **Location:** file:line214- **Severity:** Critical215- **Proof-of-concept:** `; cat /etc/passwd` or `$(whoami)` payload216- **Fix:** Use array-form exec, escape arguments, or avoid shell execution217218============================================================219PHASE 7: CSRF AND REQUEST FORGERY220============================================================221222Hunt for CSRF vulnerabilities:223224MISSING CSRF PROTECTION:225- State-changing POST/PUT/DELETE endpoints without CSRF tokens226- Missing `SameSite` cookie attribute227- CORS allowing credentials with broad origins228- Form submissions without anti-CSRF tokens229230SSRF:231- User-supplied URLs passed to HTTP clients (fetch, axios, requests)232- URL redirects without origin validation233- Webhook URLs from user input fetched server-side234- Image/preview URLs fetched without allowlist235236OPEN REDIRECTS:237- Redirect URLs from query parameters without validation238- Login redirect (`?next=` or `?redirect=`) without allowlist239- OAuth callback URLs not properly validated240241For each finding:242- **Location:** file:line243- **Severity:** Medium-High244- **Proof-of-concept:** Example malicious URL or forged request245- **Fix:** CSRF token implementation, URL validation code246247============================================================248PHASE 8: SECRETS AND CREDENTIALS249============================================================250251Hunt for hardcoded secrets:252253PATTERNS TO SEARCH:254- API keys: strings matching `[A-Za-z0-9_-]{20,}` near key-related variables255- AWS keys: `AKIA[0-9A-Z]{16}`256- Private keys: `-----BEGIN (RSA |EC |)PRIVATE KEY-----`257- JWT secrets: hardcoded strings in JWT sign/verify calls258- Database passwords: connection strings with embedded credentials259- OAuth client secrets: hardcoded in source (not env vars)260- Encryption keys: hardcoded byte arrays or strings used as keys261262CHECK CONFIGURATION:263- `.env` files committed (not in `.gitignore`)264- Secrets in `docker-compose.yml` (not using Docker secrets)265- Secrets in CI/CD config files (not using secret variables)266- Secrets in client-side/frontend code (visible to users)267268For each finding:269- **Location:** file:line270- **Severity:** Critical271- **Secret type:** [what kind of credential]272- **Fix:** Move to environment variable or secret manager273274============================================================275PHASE 9: INSECURE DESERIALIZATION276============================================================277278Hunt for deserialization attacks:279280- `JSON.parse()` on untrusted input used to construct queries or commands281- `pickle.loads()` on user-supplied data (Python)282- `yaml.load()` without `SafeLoader` (Python)283- `unserialize()` on user input (PHP)284- `ObjectInputStream` with untrusted data (Java)285- `Marshal.load()` on untrusted data (Ruby)286- XML parsing without disabling external entities (XXE)287288For each finding:289- **Location:** file:line290- **Severity:** High-Critical291- **Proof-of-concept:** Malicious serialized payload structure292- **Fix:** Safe deserialization method or input validation293294295============================================================296SELF-HEALING VALIDATION (max 2 iterations)297============================================================298299After producing the security analysis, validate thoroughness:3003011. Verify every category in the audit was actually checked (not skipped).3022. Verify every finding has a specific file:line location.3033. Verify severity ratings are justified by impact assessment.3044. Verify no false positives by re-reading flagged code in context.305306IF VALIDATION FAILS:307- Re-audit skipped categories or vague findings308- Verify or remove false positives309- Repeat up to 2 iterations310311============================================================312OUTPUT313============================================================314315## Penetration Test Report (Code Analysis)316317**Project:** [name]318**Stack:** [detected technologies]319**Scope:** [full project or specified scope]320**Date:** [date]321322### Executive Summary323324| Category | Critical | High | Medium | Low |325|----------|----------|------|--------|-----|326| Injection (SQL/NoSQL) | N | N | N | N |327| XSS | N | N | N | N |328| Auth Bypass | N | N | N | N |329| Authorization Flaws | N | N | N | N |330| Path Traversal | N | N | N | N |331| Command Injection | N | N | N | N |332| CSRF/SSRF | N | N | N | N |333| Hardcoded Secrets | N | N | N | N |334| Insecure Deserialization | N | N | N | N |335| **Total** | **N** | **N** | **N** | **N** |336337### Findings338339For each finding, in severity order:340341#### [SEVERITY] — [Finding Title]342343- **Category:** [e.g., SQL Injection]344- **Location:** `path/to/file.ts:42`345- **Description:** [What the vulnerability is]346- **Proof-of-concept:**347 ```348 [Example exploit payload or steps]349 ```350- **Impact:** [What an attacker could achieve]351- **Fix:**352 ```353 [Code showing the remediation]354 ```355356### Attack Surface Summary357- **Total routes:** N358- **Unauthenticated routes:** N359- **Routes accepting user input:** N360- **Routes with dangerous sinks:** N361362### Remediation Priority363[Ordered list: Critical findings first, with effort estimates]364365============================================================366NEXT STEPS367============================================================368369After reviewing the pentest report:370- "Fix all Critical findings immediately — these are exploitable."371- "Run `/owasp` for a systematic OWASP Top 10 compliance check."372- "Run `/encryption` to address cryptographic weaknesses."373- "Run `/dependency-scan` to fix vulnerable component findings."374- "Run `/secure` after remediation to verify the security posture improved."375376377============================================================378SELF-EVOLUTION TELEMETRY379============================================================380381After producing output, record execution metadata for the /evolve pipeline.382383Check if a project memory directory exists:384- Look for the project path in `~/.claude/projects/`385- If found, append to `skill-telemetry.md` in that memory directory386387Entry format:388```389### /pentest — {{YYYY-MM-DD}}390- Outcome: {{SUCCESS | PARTIAL | FAILED}}391- Self-healed: {{yes — what was healed | no}}392- Iterations used: {{N}} / {{N max}}393- Bottleneck: {{phase that struggled or "none"}}394- Suggestion: {{one-line improvement idea for /evolve, or "none"}}395```396397Only log if the memory directory exists. Skip silently if not found.398Keep entries concise — /evolve will parse these for skill improvement signals.399400============================================================401DO NOT402============================================================403404- Do NOT execute actual exploits — this is static code analysis only.405- Do NOT send network requests to test vulnerabilities.406- Do NOT expose full secret values — redact all but first 4 characters.407- Do NOT modify any code — report findings with suggested fixes only.408- Do NOT report theoretical vulnerabilities without code evidence — every finding must cite a specific file and line.409- Do NOT conflate "possible" with "confirmed" — clearly state confidence level.410- Do NOT skip any phase — even if the project seems small, check all categories.411- Do NOT install external tools — analyze the source code directly.