creator/Anthropic-Cybersecurity-Skills
1
0
Fork 0
mirror of https://github.com/mukul975/Anthropic-Cybersecurity-Skills.git synced 2026-07-26 14:10:59 +03:00
Code Issues Packages Projects Releases Wiki Activity

Initial commit - 611 cybersecurity skills across all subdomains

Browse Source
This commit is contained in:
mukul975
2026-02-25 10:47:44 +01:00
commit 22a7ab1462
1765 changed files with 280648 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files
skills/building-role-mining-for-rbac-optimization/SKILL.md
+235
View File
@@ -0,0 +1,235 @@
---
name: building-role-mining-for-rbac-optimization
description: Apply bottom-up and top-down role mining techniques to discover optimal RBAC roles from existing user-permission assignments, reducing role explosion and enforcing least privilege.
domain: cybersecurity
subdomain: identity-access-management
tags: [rbac, role-mining, identity-governance, access-control, least-privilege, clustering]
version: "1.0"
author: mahipal
license: MIT
---
# Building Role Mining for RBAC Optimization
## Overview
Role mining is the process of analyzing existing user-permission assignments to discover optimal roles for a Role-Based Access Control (RBAC) system. Organizations accumulate excessive permissions over time through job changes, project assignments, and ad-hoc access grants, leading to "role explosion" where thousands of granular roles exist with significant overlap. Role mining uses data analysis -- including clustering algorithms, formal concept analysis, and graph-based methods -- to consolidate permissions into a minimal set of roles that accurately represent business functions while enforcing least privilege.
## Prerequisites
- Export of current user-permission assignments (CSV/database)
- Identity governance platform or directory service access
- Python 3.9+ with pandas, scikit-learn, numpy
- Understanding of organizational structure and job functions
- Stakeholder access for role validation workshops
## Core Concepts
### Role Mining Approaches
| Approach | Description | Best For |
|----------|-------------|----------|
| Bottom-Up | Analyze existing permissions to discover common patterns | Large datasets with organic permission growth |
| Top-Down | Design roles from business requirements and job descriptions | Greenfield RBAC or organizational restructuring |
| Hybrid | Combine bottom-up analysis with top-down business validation | Most production environments |
### Role Mining Algorithms
**1. Permission Clustering**: Group users with similar permission sets using k-means or hierarchical clustering. Users in the same cluster share a common role.
**2. Formal Concept Analysis (FCA)**: Mathematical framework that identifies complete set of concepts (user groups sharing exact permission sets) from a binary user-permission matrix.
**3. Graph-Based Mining**: Model users and permissions as a bipartite graph, then find dense subgraphs representing candidate roles.
**4. Boolean Matrix Decomposition**: Decompose the user-permission matrix U into U ≈ R × P where R maps users to roles and P maps roles to permissions.
### Role Mining Metrics
| Metric | Formula | Target |
|--------|---------|--------|
| Role Count | Total distinct roles after mining | Minimize |
| Coverage | Permissions explained by mined roles / Total permissions | > 95% |
| Weighted Structural Complexity (WSC) | Sum of role-user + role-permission assignments | Minimize |
| Deviation | Extra permissions not covered by assigned roles | < 5% |
## Implementation Steps
### Step 1: Extract User-Permission Data
Collect the current access state from all identity sources:
```python
import pandas as pd
import numpy as np
# Load user-permission assignments
# Format: user_id, permission_id (one row per assignment)
assignments = pd.read_csv("user_permissions.csv")
# Create binary user-permission matrix (UPA matrix)
upa_matrix = assignments.pivot_table(
index="user_id",
columns="permission_id",
aggfunc="size",
fill_value=0
)
upa_matrix = (upa_matrix > 0).astype(int)
print(f"Users: {upa_matrix.shape[0]}")
print(f"Permissions: {upa_matrix.shape[1]}")
print(f"Assignments: {assignments.shape[0]}")
print(f"Density: {upa_matrix.values.sum() / upa_matrix.size:.2%}")
```
### Step 2: Bottom-Up Role Discovery Using Clustering
```python
from sklearn.cluster import AgglomerativeClustering
from sklearn.metrics import silhouette_score
def find_optimal_clusters(matrix, max_k=50):
"""Find optimal number of roles using silhouette analysis."""
scores = []
for k in range(2, min(max_k, matrix.shape[0])):
clustering = AgglomerativeClustering(
n_clusters=k, metric="jaccard", linkage="average"
)
labels = clustering.fit_predict(matrix)
score = silhouette_score(matrix, labels, metric="jaccard")
scores.append((k, score))
optimal_k = max(scores, key=lambda x: x[1])[0]
return optimal_k, scores
def mine_roles_clustering(upa_matrix, n_clusters):
"""Mine roles using hierarchical clustering on Jaccard distance."""
clustering = AgglomerativeClustering(
n_clusters=n_clusters, metric="jaccard", linkage="average"
)
user_matrix = upa_matrix.values
labels = clustering.fit_predict(user_matrix)
roles = {}
for cluster_id in range(n_clusters):
cluster_users = upa_matrix.index[labels == cluster_id]
cluster_permissions = upa_matrix.loc[cluster_users]
# Core role = permissions held by >80% of cluster members
permission_frequency = cluster_permissions.mean()
core_permissions = permission_frequency[permission_frequency >= 0.8].index.tolist()
roles[f"Role_{cluster_id}"] = {
"permissions": core_permissions,
"user_count": len(cluster_users),
"users": cluster_users.tolist(),
"coverage": permission_frequency[permission_frequency >= 0.8].mean()
}
return roles, labels
```
### Step 3: Formal Concept Analysis
```python
def mine_roles_fca(upa_matrix, min_support=3):
"""Mine roles using Formal Concept Analysis (frequent closed itemsets)."""
from itertools import combinations
users = upa_matrix.index.tolist()
permissions = upa_matrix.columns.tolist()
concepts = []
# Find all maximal permission sets shared by at least min_support users
for size in range(len(permissions), 0, -1):
for perm_combo in combinations(permissions, size):
perm_set = set(perm_combo)
# Find users who have ALL permissions in this set
matching_users = []
for user in users:
user_perms = set(upa_matrix.columns[upa_matrix.loc[user] == 1])
if perm_set.issubset(user_perms):
matching_users.append(user)
if len(matching_users) >= min_support:
# Check if this is a closed concept (no superset with same extent)
is_closed = True
for concept in concepts:
if set(matching_users) == set(concept["users"]) and \
perm_set.issubset(set(concept["permissions"])):
is_closed = False
break
if is_closed:
concepts.append({
"permissions": list(perm_set),
"users": matching_users,
"support": len(matching_users)
})
if len(concepts) > 100: # Limit for performance
break
return concepts
```
### Step 4: Evaluate and Select Roles
```python
def evaluate_role_set(roles, upa_matrix):
"""Evaluate the quality of a mined role set."""
total_assignments = upa_matrix.values.sum()
covered_assignments = 0
extra_assignments = 0
for role_name, role_data in roles.items():
role_perms = set(role_data["permissions"])
for user in role_data["users"]:
user_perms = set(upa_matrix.columns[upa_matrix.loc[user] == 1])
covered = role_perms.intersection(user_perms)
extra = role_perms - user_perms
covered_assignments += len(covered)
extra_assignments += len(extra)
metrics = {
"total_roles": len(roles),
"total_assignments": total_assignments,
"covered_assignments": covered_assignments,
"coverage_rate": covered_assignments / total_assignments if total_assignments else 0,
"extra_permissions": extra_assignments,
"deviation_rate": extra_assignments / (covered_assignments + extra_assignments) if (covered_assignments + extra_assignments) else 0,
"avg_role_size": np.mean([len(r["permissions"]) for r in roles.values()]),
"avg_users_per_role": np.mean([r["user_count"] for r in roles.values()]),
}
return metrics
```
### Step 5: Business Validation
After mining candidate roles:
1. Map mined roles to business functions (department, job title)
2. Conduct workshops with business unit managers to validate role definitions
3. Identify outlier permissions that indicate misconfiguration
4. Refine roles based on feedback and re-evaluate metrics
5. Document role definitions with business justification
## Validation Checklist
- [ ] User-permission matrix extracted from all identity sources
- [ ] Multiple mining algorithms compared (clustering, FCA)
- [ ] Optimal role count determined via silhouette analysis or WSC
- [ ] Coverage rate exceeds 95% of existing assignments
- [ ] Deviation rate below 5% (minimal extra permissions)
- [ ] Mined roles validated with business stakeholders
- [ ] Role hierarchy defined (parent-child inheritance)
- [ ] Exception/outlier permissions documented
- [ ] Migration plan created for transitioning to new role model
- [ ] Ongoing role governance process defined
## References
- [Role Mining: Optimizing RBAC - NIST](https://csrc.nist.gov/projects/role-based-access-control)
- [RBAC Standard - ANSI/INCITS 359-2012](https://www.incits.org/)
- [Formal Concept Analysis for Role Engineering](https://link.springer.com/chapter/10.1007/978-3-540-73070-6_7)
- [scikit-learn Clustering Documentation](https://scikit-learn.org/stable/modules/clustering.html)
skills/building-role-mining-for-rbac-optimization/assets/template.md
+52
View File
@@ -0,0 +1,52 @@
# Role Mining Project Template
## Project Overview
| Field | Value |
|-------|-------|
| Organization | |
| Project Lead | |
| Start Date | |
| Target Completion | |
| Identity Sources | AD / Azure / AWS / Applications |
## Data Collection Summary
| Source | Users | Permissions | Assignments |
|--------|-------|-------------|-------------|
| Active Directory | | | |
| AWS IAM | | | |
| Azure AD | | | |
| Applications | | | |
| **Total (Deduplicated)** | | | |
## Mining Results
| Algorithm | Roles Found | Coverage | Deviation | WSC |
|-----------|-------------|----------|-----------|-----|
| Clustering (k=___) | | | | |
| Intersection Mining | | | | |
| Selected Approach | | | | |
## Proposed Role Definitions
| Role Name | Department | Permissions | Users | Status |
|-----------|------------|-------------|-------|--------|
| | | | | Draft/Validated/Approved |
| | | | | |
## Stakeholder Validation
| Business Unit | Reviewer | Roles Reviewed | Approved | Date |
|--------------|----------|---------------|----------|------|
| | | | Yes/No | |
## Migration Plan
- [ ] Roles created in identity governance platform
- [ ] User-role assignments configured
- [ ] Individual permission grants removed
- [ ] Test users validated access
- [ ] Full migration completed
- [ ] Post-migration access verification
- [ ] Old permissions cleanup confirmed
skills/building-role-mining-for-rbac-optimization/references/standards.md
+46
View File
@@ -0,0 +1,46 @@
# Role Mining for RBAC Optimization - Standards Reference
## RBAC Standards
### ANSI/INCITS 359-2012 - Core RBAC
- Defines User, Role, Permission, Session abstractions
- Role assignment: users are assigned to roles
- Permission assignment: permissions are assigned to roles
- Role hierarchy: senior roles inherit junior role permissions
- Separation of Duty constraints (static and dynamic)
### NIST RBAC Model (SP 800-162)
- Core RBAC: Basic user-role and role-permission mappings
- Hierarchical RBAC: Role inheritance relationships
- Constrained RBAC: Static and dynamic separation of duties
- Symmetric RBAC: Combined user-centric and permission-centric views
## Identity Governance Standards
### ISO 27001:2022 - A.5.15 Access Control
- Access control policy based on business and security requirements
- Roles determined by job function
- Regular review of access rights
- Formal authorization for privilege changes
### NIST SP 800-53 Rev 5
- AC-2: Account Management
- AC-3: Access Enforcement
- AC-5: Separation of Duties
- AC-6: Least Privilege
- AC-16: Security and Privacy Attributes
- AC-24: Access Control Decisions
## Role Mining Research
### Key Algorithms
- **RoleMiner (Vaidya et al., 2007)**: Iterative role mining minimizing WSC
- **CompleteMiner / FastMiner (Vaidya et al., 2006)**: Complete vs. approximate algorithms
- **ORCA (Schlegelmilch & Steffens, 2005)**: Clustering-based approach
- **Graph Optimization (Lu et al., 2008)**: Graph-based role mining
### Quality Metrics
- Weighted Structural Complexity: min(|UA| + |PA| + |Roles|)
- Boolean Matrix Decomposition error
- Jaccard similarity between mined and original access
- Role coverage percentage
skills/building-role-mining-for-rbac-optimization/references/workflows.md
+82
View File
@@ -0,0 +1,82 @@
# Role Mining for RBAC Optimization - Workflows
## End-to-End Role Mining Workflow
```
Phase 1: DATA COLLECTION (Week 1-2)
├── Export user-permission data from all identity sources
│ ├── Active Directory group memberships
│ ├── Cloud IAM role assignments
│ ├── Application-level permissions
│ └── Database access grants
├── Collect HR data (job titles, departments, cost centers)
├── Normalize data into User-Permission Assignment (UPA) matrix
└── Clean data: remove disabled accounts, system accounts
Phase 2: ANALYSIS (Week 3-4)
├── Run clustering algorithms (hierarchical, k-means)
├── Run Formal Concept Analysis for exact role candidates
├── Compare results using WSC and coverage metrics
├── Identify optimal number of roles via silhouette analysis
└── Map candidate roles to organizational structure
Phase 3: VALIDATION (Week 5-6)
├── Present candidate roles to business unit managers
├── Validate each role against job descriptions
├── Identify and resolve outlier permissions
├── Define role hierarchy (inheritance relationships)
└── Agree on role names and descriptions
Phase 4: IMPLEMENTATION (Week 7-8)
├── Create roles in identity governance platform
├── Assign users to validated roles
├── Remove individual permission assignments
├── Test access for sample users in each role
└── Document role definitions and approval chain
Phase 5: GOVERNANCE (Ongoing)
├── Monitor for permission drift
├── Quarterly role effectiveness review
├── Re-run mining annually to detect new patterns
└── Track role count and WSC metrics over time
```
## Data Normalization Workflow
```
Raw Data Sources
│
├── AD: user → group → permissions
│ Normalize to: user_id, permission_id
│
├── AWS: user/role → policy → actions
│ Normalize to: user_id, permission_id
│
├── Azure: user → role → permissions
│ Normalize to: user_id, permission_id
│
└── Applications: user → app_role → features
Normalize to: user_id, permission_id
Merge all sources → Deduplicate → Create UPA matrix
```
## Role Consolidation Workflow
```
Mining produces N candidate roles
│
├── Remove roles with < 3 users (outliers)
│
├── Merge roles with > 90% Jaccard similarity
│
├── Identify hierarchical relationships:
│ └── If Role A permissions ⊂ Role B permissions
│ → Role A is junior to Role B
│
├── Check for SoD violations:
│ └── Does any role combine conflicting permissions?
│ → Split into separate roles if needed
│
└── Final role set with hierarchy and constraints
```
skills/building-role-mining-for-rbac-optimization/scripts/process.py
+251
View File
@@ -0,0 +1,251 @@
#!/usr/bin/env python3
"""
Role Mining Engine for RBAC Optimization
Implements multiple role mining algorithms (clustering, FCA) on user-permission
assignment data to discover optimal RBAC roles. Generates role definitions,
coverage reports, and migration plans.
Requirements:
pip install pandas numpy scikit-learn
"""
import csv
import json
from collections import defaultdict
from itertools import combinations
from pathlib import Path
import numpy as np
import pandas as pd
from sklearn.cluster import AgglomerativeClustering
from sklearn.metrics import silhouette_score
class RoleMiningEngine:
"""Core role mining engine supporting multiple algorithms."""
def __init__(self, assignments_file=None):
self.upa_matrix = None
self.user_metadata = {}
self.mined_roles = {}
if assignments_file:
self.load_assignments(assignments_file)
def load_assignments(self, filepath):
"""Load user-permission assignments from CSV (user_id, permission_id)."""
df = pd.read_csv(filepath)
required = {"user_id", "permission_id"}
if not required.issubset(df.columns):
raise ValueError(f"CSV must contain columns: {required}")
self.upa_matrix = df.pivot_table(
index="user_id", columns="permission_id",
aggfunc="size", fill_value=0
)
self.upa_matrix = (self.upa_matrix > 0).astype(int)
print(f"[OK] Loaded UPA matrix: {self.upa_matrix.shape[0]} users x "
f"{self.upa_matrix.shape[1]} permissions")
print(f" Total assignments: {self.upa_matrix.values.sum()}")
density = self.upa_matrix.values.sum() / self.upa_matrix.size
print(f" Matrix density: {density:.2%}")
def load_user_metadata(self, filepath):
"""Load user HR data (user_id, department, title, location)."""
df = pd.read_csv(filepath)
for _, row in df.iterrows():
self.user_metadata[row["user_id"]] = row.to_dict()
def find_optimal_k(self, max_k=50):
"""Determine optimal number of roles using silhouette analysis."""
if self.upa_matrix is None:
raise ValueError("No data loaded")
matrix = self.upa_matrix.values
max_k = min(max_k, matrix.shape[0] - 1)
scores = []
for k in range(2, max_k + 1):
clustering = AgglomerativeClustering(
n_clusters=k, metric="jaccard", linkage="average"
)
labels = clustering.fit_predict(matrix)
score = silhouette_score(matrix, labels, metric="jaccard")
scores.append({"k": k, "silhouette": round(score, 4)})
best = max(scores, key=lambda x: x["silhouette"])
print(f"[OK] Optimal k={best['k']} (silhouette={best['silhouette']})")
return best["k"], scores
def mine_roles_clustering(self, n_clusters=None, threshold=0.8):
"""Mine roles using hierarchical clustering with Jaccard distance."""
if self.upa_matrix is None:
raise ValueError("No data loaded")
if n_clusters is None:
n_clusters, _ = self.find_optimal_k()
matrix = self.upa_matrix.values
clustering = AgglomerativeClustering(
n_clusters=n_clusters, metric="jaccard", linkage="average"
)
labels = clustering.fit_predict(matrix)
roles = {}
for cluster_id in range(n_clusters):
mask = labels == cluster_id
cluster_users = self.upa_matrix.index[mask].tolist()
cluster_data = self.upa_matrix.loc[cluster_users]
perm_freq = cluster_data.mean()
core_perms = perm_freq[perm_freq >= threshold].index.tolist()
# Determine role name from user metadata
role_label = f"Role_{cluster_id:03d}"
if self.user_metadata:
depts = [self.user_metadata.get(u, {}).get("department", "Unknown")
for u in cluster_users]
dept_counts = defaultdict(int)
for d in depts:
dept_counts[d] += 1
if dept_counts:
dominant_dept = max(dept_counts, key=dept_counts.get)
role_label = f"{dominant_dept}_Role_{cluster_id:03d}"
roles[role_label] = {
"permissions": core_perms,
"user_count": len(cluster_users),
"users": cluster_users,
"permission_count": len(core_perms),
}
self.mined_roles = roles
print(f"[OK] Mined {len(roles)} roles via clustering")
return roles
def mine_roles_intersection(self, min_users=3):
"""Mine roles by finding common permission intersections."""
if self.upa_matrix is None:
raise ValueError("No data loaded")
user_perm_sets = {}
for user in self.upa_matrix.index:
perms = set(self.upa_matrix.columns[self.upa_matrix.loc[user] == 1])
user_perm_sets[user] = perms
# Find unique permission sets shared by multiple users
perm_set_users = defaultdict(list)
for user, perms in user_perm_sets.items():
key = frozenset(perms)
perm_set_users[key].append(user)
roles = {}
role_idx = 0
for perm_set, users in perm_set_users.items():
if len(users) >= min_users:
roles[f"ExactRole_{role_idx:03d}"] = {
"permissions": sorted(perm_set),
"user_count": len(users),
"users": users,
"permission_count": len(perm_set),
}
role_idx += 1
self.mined_roles = roles
print(f"[OK] Mined {len(roles)} exact-match roles "
f"(min {min_users} users per role)")
return roles
def evaluate_roles(self, roles=None):
"""Calculate quality metrics for a set of mined roles."""
if roles is None:
roles = self.mined_roles
if not roles:
return {"error": "No roles to evaluate"}
total_assignments = int(self.upa_matrix.values.sum())
covered = 0
extra = 0
for role_data in roles.values():
role_perms = set(role_data["permissions"])
for user in role_data["users"]:
user_perms = set(
self.upa_matrix.columns[self.upa_matrix.loc[user] == 1]
)
covered += len(role_perms & user_perms)
extra += len(role_perms - user_perms)
total_role_assignments = sum(
r["user_count"] + r["permission_count"] for r in roles.values()
)
metrics = {
"total_roles": len(roles),
"total_original_assignments": total_assignments,
"covered_assignments": covered,
"extra_permissions_granted": extra,
"coverage_rate": round(covered / total_assignments, 4) if total_assignments else 0,
"deviation_rate": round(extra / (covered + extra), 4) if (covered + extra) else 0,
"wsc": total_role_assignments + len(roles),
"avg_permissions_per_role": round(
np.mean([r["permission_count"] for r in roles.values()]), 1
),
"avg_users_per_role": round(
np.mean([r["user_count"] for r in roles.values()]), 1
),
}
return metrics
def export_roles(self, output_path):
"""Export mined roles to JSON for import into IGA platform."""
export = {
"generated_at": pd.Timestamp.now().isoformat(),
"metrics": self.evaluate_roles(),
"roles": {}
}
for name, data in self.mined_roles.items():
export["roles"][name] = {
"name": name,
"permissions": data["permissions"],
"user_count": data["user_count"],
"permission_count": data["permission_count"],
}
with open(output_path, "w") as f:
json.dump(export, f, indent=2)
print(f"[OK] Exported {len(self.mined_roles)} roles to {output_path}")
def generate_migration_plan(self, output_path):
"""Generate a CSV migration plan mapping users to new roles."""
rows = []
for role_name, role_data in self.mined_roles.items():
for user in role_data["users"]:
rows.append({
"user_id": user,
"new_role": role_name,
"permissions_in_role": len(role_data["permissions"]),
"current_permissions": int(self.upa_matrix.loc[user].sum()),
})
df = pd.DataFrame(rows)
df.to_csv(output_path, index=False)
print(f"[OK] Migration plan exported to {output_path}")
if __name__ == "__main__":
print("=" * 60)
print("Role Mining Engine for RBAC Optimization")
print("=" * 60)
print()
print("Usage:")
print(" engine = RoleMiningEngine('user_permissions.csv')")
print(" engine.load_user_metadata('hr_data.csv')")
print(" optimal_k, scores = engine.find_optimal_k()")
print(" roles = engine.mine_roles_clustering(n_clusters=optimal_k)")
print(" metrics = engine.evaluate_roles()")
print(" engine.export_roles('mined_roles.json')")
print(" engine.generate_migration_plan('migration_plan.csv')")