feat: full scan mode, init command, 18 gates, and 100% test coverage

.controlgate.yml

@@ -4,14 +4,24 @@ gov: false                      # If true, evaluates against FedRAMP baselines i
 catalog: baseline/nist80053r5_full_catalog_enriched.json
 
 gates:
-  secrets:    { enabled: true,  action: block }
-  crypto:     { enabled: true,  action: block }
-  iam:        { enabled: true,  action: warn  }
-  sbom:       { enabled: true,  action: warn  }
-  iac:        { enabled: true,  action: block }
-  input:      { enabled: true,  action: block }
-  audit:      { enabled: true,  action: warn  }
-  change:     { enabled: true,  action: warn  }
+  secrets:             { enabled: true,  action: block }
+  crypto:              { enabled: true,  action: block }
+  iam:                 { enabled: true,  action: warn  }
+  sbom:                { enabled: true,  action: warn  }
+  iac:                 { enabled: true,  action: block }
+  input_validation:    { enabled: true,  action: block }
+  audit:               { enabled: true,  action: warn  }
+  change_control:      { enabled: true,  action: warn  }
+  deps:                { enabled: true,  action: warn  }
+  api:                 { enabled: true,  action: warn  }
+  privacy:             { enabled: true,  action: warn  }
+  resilience:          { enabled: true,  action: warn  }
+  incident:            { enabled: true,  action: warn  }
+  observability:       { enabled: true,  action: warn  }
+  memsafe:             { enabled: true,  action: warn  }
+  license:             { enabled: true,  action: warn  }
+  aiml:                { enabled: true,  action: warn  }
+  container:           { enabled: true,  action: warn  }
 
 thresholds:
   block_on:   [CRITICAL, HIGH]  # Block if finding severity matches

.pre-commit-config.yaml

@@ -23,3 +23,13 @@ repos:
       - id: mypy
         additional_dependencies: [ types-PyYAML, pytest ]
         args: [ --config-file, pyproject.toml ]
+
+  - repo: local
+    hooks:
+      - id: bump-version
+        name: Auto-bump version if unchanged from main
+        entry: python hooks/bump_version.py
+        language: python
+        always_run: true
+        pass_filenames: false
+        stages: [pre-commit]

CHANGELOG.md

@@ -8,6 +8,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 ## [Unreleased]
 
 ### Added
+- **10 New Security Gates**: Expanded scanning capabilities to a total of 18 gates, including API, Privacy, Resilience, Incident, Observability, MemSafe, License, AI/ML, Container, and Dependency vulnerabilities.
 - **FedRAMP Support**: ControlGate can now evaluate code changes against FedRAMP baselines (LI-SaaS, Low, Moderate, High) via the `--gov` CLI flag or `gov: true` in `.controlgate.yml`. This integrates directly with the enriched `fedramp_membership` catalog data.
 
 ## [0.1.3] - 2026-02-20

README.md

@@ -14,6 +14,12 @@ ControlGate is an AI-powered agent skill that scans your code changes against th
 # Install
 pip install controlgate
 
+# Bootstrap project (creates .controlgate.yml, pre-commit hook, CI workflows)
+controlgate init
+
+# Full repo audit (first-time scan)
+controlgate scan --mode full --format markdown
+
 # Scan staged changes (NIST baseline)
 controlgate scan --mode pre-commit --format markdown
 
@@ -33,12 +39,12 @@ git commit / Pull Request
        ↓
 ControlGate intercepts the diff
        ↓
-8 Security Gates scan against 370 non-negotiable NIST controls
+18 Security Gates scan against 370 non-negotiable NIST controls
        ↓
 Verdict: BLOCK 🚫 / WARN ⚠️ / PASS ✅
 ```
 
-## The Eight Security Gates
+## The 18 Security Gates
 
 | # | Gate | NIST Families | What It Catches |
 |---|------|---------------|-----------------|
@@ -47,9 +53,19 @@ Verdict: BLOCK 🚫 / WARN ⚠️ / PASS ✅
 | 3 | 🛡️ IAM | AC-3, AC-5, AC-6 | Wildcard IAM, missing auth, overprivileged roles |
 | 4 | 📦 Supply Chain | SR-3, SR-11, SA-10 | Unpinned deps, missing lockfiles, build tampering |
 | 5 | 🏗️ IaC | CM-2, CM-6, SC-7 | Public buckets, `0.0.0.0/0` rules, root containers |
-| 6 | ✅ Input | SI-10, SI-11 | SQL injection, `eval()`, exposed stack traces |
+| 6 | ✅ Input Validation | SI-10, SI-11 | SQL injection, `eval()`, exposed stack traces |
 | 7 | 📋 Audit | AU-2, AU-3, AU-12 | Missing security logging, PII in logs |
-| 8 | 🔄 Change | CM-3, CM-4, CM-5 | Unauthorized config changes, missing CODEOWNERS |
+| 8 | 🔄 Change Control | CM-3, CM-4, CM-5 | Unauthorized config changes, missing CODEOWNERS |
+| 9 | 🧩 Dependencies | SA-12, RA-5, SI-2 | Vulnerable packages, missing lockfiles |
+| 10 | 🌐 API Security | SC-8, AC-3, SI-10 | Unauthenticated endpoints, missing rate limiting |
+| 11 | 🔐 Privacy | AR-2, DM-1, IP-1 | PII exposure, missing data classification |
+| 12 | 🔁 Resilience | CP-2, CP-10, SC-5 | Missing retry logic, no circuit breakers |
+| 13 | 🚨 Incident Response | IR-2, IR-4, IR-6 | Missing error handlers, no alerting integration |
+| 14 | 👁️ Observability | AU-6, SI-4, CA-7 | Missing health checks, no structured logging |
+| 15 | 🧠 Memory Safety | SI-16, SA-11, SA-15 | Buffer overflows, unsafe memory operations |
+| 16 | ⚖️ License Compliance | SA-4, SR-3 | GPL contamination, unlicensed dependencies |
+| 17 | 🤖 AI/ML Security | SA-11, SI-7, AC-3 | Untrusted model sources, prompt injection risk |
+| 18 | 🐳 Container Security | CM-6, AC-6, SC-7 | Root containers, privileged mode, `latest` tags |
 
 ## Installation
 
@@ -87,17 +103,26 @@ Create a `.controlgate.yml` in your project root:
 ```yaml
 baseline: moderate              # low | moderate | high | privacy | li-saas
 gov: false                      # set to true to evaluate against FedRAMP baselines
-catalog: baseline/nist80053r5_full_catalog_enriched.json
 
 gates:
-  secrets:    { enabled: true,  action: block }
-  crypto:     { enabled: true,  action: block }
-  iam:        { enabled: true,  action: warn  }
-  sbom:       { enabled: true,  action: warn  }
-  iac:        { enabled: true,  action: block }
-  input:      { enabled: true,  action: block }
-  audit:      { enabled: true,  action: warn  }
-  change:     { enabled: true,  action: warn  }
+  secrets:          { enabled: true,  action: block }
+  crypto:           { enabled: true,  action: block }
+  iam:              { enabled: true,  action: warn  }
+  sbom:             { enabled: true,  action: warn  }
+  iac:              { enabled: true,  action: block }
+  input_validation: { enabled: true,  action: block }
+  audit:            { enabled: true,  action: warn  }
+  change_control:   { enabled: true,  action: warn  }
+  deps:             { enabled: true,  action: warn  }
+  api:              { enabled: true,  action: warn  }
+  privacy:          { enabled: true,  action: warn  }
+  resilience:       { enabled: true,  action: warn  }
+  incident:         { enabled: true,  action: warn  }
+  observability:    { enabled: true,  action: warn  }
+  memsafe:          { enabled: true,  action: warn  }
+  license:          { enabled: true,  action: warn  }
+  aiml:             { enabled: true,  action: warn  }
+  container:        { enabled: true,  action: warn  }
 
 thresholds:
   block_on:   [CRITICAL, HIGH]
@@ -106,16 +131,25 @@ thresholds:
 
 exclusions:
   paths: ["tests/**", "docs/**", "*.md"]
+
+full_scan:
+  extensions: [.py, .js, .ts, .go, .tf, .yaml, .yml, .json, .sh]
+  skip_dirs: [.git, node_modules, .venv, dist, build]
 ```
 
 ## CLI Usage
 
 ```bash
+# Bootstrap project
+controlgate init
+controlgate init --baseline high --no-docs
+
 # Scan staged changes (pre-commit mode)
 controlgate scan --mode pre-commit --format markdown
 
-# Scan explicitly against FedRAMP baselines (pre-commit mode)
-controlgate scan --gov --baseline moderate --mode pre-commit --format markdown
+# Full repository scan
+controlgate scan --mode full --format markdown
+controlgate scan --mode full --path ./src --format json
 
 # Scan PR diff
 controlgate scan --mode pr --target-branch main --format json markdown sarif
@@ -126,11 +160,12 @@ controlgate scan --gov --baseline high --mode pr --target-branch main --format j
 # Scan a saved diff file
 controlgate scan --diff-file path/to/diff --format json
 
-# Scan a saved diff file against FedRAMP baselines
-controlgate scan --gov --baseline li-saas --diff-file path/to/diff --format json
-
 # Output reports to directory
 controlgate scan --output-dir .controlgate/reports --format json markdown sarif
+
+# Catalog management
+controlgate update-catalog
+controlgate catalog-info
 ```
 
 ## Output Formats
@@ -168,7 +203,7 @@ The agent prompts and workflows are located in the [`skills/`](skills/) director
 ## Test Suite
 
 To validate the capabilities of ControlGate, we maintain a standalone test suite at [sadayamuthu/controlgate-test-suite](https://github.com/sadayamuthu/controlgate-test-suite).
-This suite contains intentionally vulnerable projects spanning multiple languages and frameworks, specifically designed to trigger each of the 8 Security Gates. It is automatically executed in the CI pipeline to ensure zero regression in detection capabilities.
+This suite contains intentionally vulnerable projects spanning multiple languages and frameworks, specifically designed to trigger each of the 18 Security Gates. It is automatically executed in the CI pipeline to ensure zero regression in detection capabilities.
 
 ## Data Source
 

docs/gates/gate-01-secrets/design.md

@@ -0,0 +1,58 @@
+# Gate 1 — Secrets & Credential Gate
+
+**gate_id:** `secrets`
+**NIST Controls:** IA-5, IA-6, SC-12, SC-28
+**Priority:** 🔴 High
+
+---
+
+## Purpose
+
+Prevents hardcoded secrets, API keys, tokens, and private keys from being committed to source control. Credentials committed to git are permanently exposed in history, even after removal, and are a leading cause of cloud account compromise. This gate provides an automated pre-commit tripwire for the most common credential formats used across AWS, GCP, GitHub, OpenAI, Stripe, and database connection strings.
+
+---
+
+## What This Gate Detects
+
+| Detection | Why It Matters | NIST Control |
+|---|---|---|
+| AWS Access Key ID (AKIA/ASIA format) | Direct cloud account takeover | IA-5 |
+| AWS Secret Access Key (40-char base64) | Paired with key ID enables full AWS API access | IA-5 |
+| Google API Key (AIza prefix) | Unauthorized GCP service usage and billing fraud | IA-5 |
+| Hardcoded credential assignment (`password =`, `token =`, etc.) | Generic pattern catches language-agnostic credential leaks | SC-28 |
+| Private key files (RSA, EC, DSA, OpenSSH) | Key compromise enables impersonation and decryption of historical traffic | SC-12 |
+| X.509 certificate files | Certificates in repos risk unauthorized service impersonation | SC-12 |
+| GitHub Personal Access Tokens (ghp_ prefix) | Repo access, Actions secrets, and org data at risk | IA-5 |
+| API secret keys (sk- prefix) | OpenAI/Stripe-style keys enabling fraudulent API usage | IA-5 |
+| Bearer tokens in source code | Runtime tokens should not appear in code | IA-6 |
+| Database connection strings with embedded credentials | Leaks DB host, port, username, and password | SC-28 |
+| High-entropy quoted strings ≥20 chars | Catches randomized secrets not matching known formats | IA-5 |
+| Sensitive file types committed (.env, .pem, .key, .p12, .pfx, .jks) | Entire file classes should never appear in source control | SC-28 |
+
+---
+
+## Scope
+
+- **Scans:** all added lines in every file in the diff
+- **File types targeted:** all files; additionally performs file-path checks for sensitive file extensions
+- **Special detection:** Shannon entropy analysis — any quoted string ≥20 characters with entropy ≥4.5 bits/char is flagged even if it doesn't match a known pattern. Evidence is truncated to 120 characters.
+
+---
+
+## Known Limitations
+
+- Does not scan deleted or unmodified lines
+- Entropy detection produces false positives on long base64-encoded non-secret values (e.g., public keys, checksums)
+- Known secret formats are a subset of real-world credential types; novel formats require new patterns
+- Will not detect secrets stored in binary files
+
+---
+
+## NIST Control Mapping
+
+| Control ID | Title | How This Gate Addresses It |
+|---|---|---|
+| IA-5 | Authenticator Management | Detects hardcoded passwords, tokens, and API keys that should be managed through a secrets manager |
+| IA-6 | Authentication Feedback | Detects bearer tokens embedded in source code where they may be logged or exposed |
+| SC-12 | Cryptographic Key Establishment and Management | Detects private keys and certificates committed to repositories |
+| SC-28 | Protection of Information at Rest | Detects credentials and sensitive file types that expose data-at-rest protection keys |

docs/gates/gate-01-secrets/implementation.md

@@ -0,0 +1,73 @@
+# Gate 1 — Secrets & Credential Gate: Implementation Reference
+
+**Source file:** `src/controlgate/gates/secrets_gate.py`
+**Test file:** `tests/test_gates/test_secrets_gate.py`
+**Class:** `SecretsGate`
+**gate_id:** `secrets`
+**mapped_control_ids:** `["IA-5", "IA-6", "SC-12", "SC-28"]`
+
+---
+
+## Scan Method
+
+`scan()` iterates every `diff_file` and calls two sub-methods:
+1. `_check_sensitive_file()` — matches the file path against `_SENSITIVE_FILE_PATTERNS`; fires one SC-28 finding per file if the path matches (breaks after first match per file)
+2. `_check_line()` — for each added line, runs all `_PATTERNS` regex matches, then runs Shannon entropy analysis on any quoted string ≥20 chars
+
+---
+
+## Patterns
+
+| # | Regex | Description | Control | Remediation |
+|---|---|---|---|---|
+| 1 | `(?:AKIA\|ASIA)[0-9A-Z]{16}` | AWS Access Key ID detected | IA-5 | Use IAM roles or AWS Secrets Manager instead of hardcoded keys |
+| 2 | `(?:"\|')(?:[A-Za-z0-9/+=]{40})(?:"\|')` | Possible AWS Secret Access Key detected | IA-5 | Use IAM roles or AWS Secrets Manager instead of hardcoded keys |
+| 3 | `AIza[0-9A-Za-z\-_]{35}` | Google API Key detected | IA-5 | Use GCP Secret Manager or environment variables |
+| 4 | `(?i)(?:password\|passwd\|pwd\|secret\|token\|api[_-]?key\|auth[_-]?token\|access[_-]?token)\s*[:=]\s*["'][^"']{4,}["']` | Hardcoded credential detected | SC-28 | Move to environment variable or secrets manager (AWS SSM, GCP Secret Manager, Azure Key Vault) |
+| 5 | `(?i)(?:password\|passwd\|pwd\|secret\|token\|api[_-]?key)\s*=\s*(?!None\|null\|""\|''\|os\.environ\|env\(\|getenv)[^\s#]{4,}` | Hardcoded credential in assignment | SC-28 | Move to environment variable or secrets manager |
+| 6 | `-----BEGIN (?:RSA \|EC \|DSA \|OPENSSH )?PRIVATE KEY-----` | Private key committed to repository | SC-12 | Never commit private keys. Use a secrets manager or secure key storage |
+| 7 | `-----BEGIN CERTIFICATE-----` | Certificate file committed to repository | SC-12 | Manage certificates through a PKI or secrets manager, not source control |
+| 8 | `ghp_[0-9a-zA-Z]{36}` | GitHub Personal Access Token detected | IA-5 | Use GitHub Apps or GITHUB_TOKEN instead of personal access tokens |
+| 9 | `sk-[0-9a-zA-Z]{20,}` | API secret key detected (OpenAI/Stripe pattern) | IA-5 | Use environment variables or a secrets manager for API keys |
+| 10 | `(?i)bearer\s+[a-z0-9\-._~+/]+=*` | Bearer token detected in source code | IA-6 | Tokens should be loaded from environment or config, not hardcoded |
+| 11 | `(?i)(?:mongodb\|postgres(?:ql)?\|mysql\|redis\|amqp)://[^\s:]+:[^\s@]+@` | Database connection string with embedded credentials | SC-28 | Use environment variables for connection strings with credentials |
+
+**Sensitive file path patterns** (checked against `diff_file.path`, fires SC-28):
+
+| Pattern | Matches |
+|---|---|
+| `\.env(?:\..+)?$` | `.env`, `.env.prod`, `.env.local`, etc. |
+| `(?i)credentials` | Any file with "credentials" in name |
+| `(?i)\.pem$` | PEM certificate/key files |
+| `(?i)\.key$` | Key files |
+| `(?i)\.p12$` | PKCS#12 key stores |
+| `(?i)\.pfx$` | Personal Information Exchange files |
+| `(?i)\.jks$` | Java KeyStore files |
+
+---
+
+## Special Detection Logic
+
+**Shannon entropy analysis** — runs after the pattern loop in `_check_line()`:
+
+1. Scans the line for quoted strings matching `["']([A-Za-z0-9+/=\-_]{20,})["']`
+2. For each candidate token of length ≥20: computes Shannon entropy (bits per character)
+3. If entropy ≥ 4.5 AND no finding from this gate already exists for that line number, fires an IA-5 finding
+4. The duplicate-check prevents double-firing when a pattern already matched the same line
+
+Constants: `_ENTROPY_THRESHOLD = 4.5`, `_MIN_LENGTH_FOR_ENTROPY = 20`
+
+---
+
+## Test Coverage
+
+| Test | What It Verifies |
+|---|---|
+| `test_detects_aws_key` | AKIA/ASIA format AWS access key triggers an IA-5 finding with "AWS" in the description |
+| `test_detects_hardcoded_password` | `DB_PASSWORD = "super_secret_123"` assignment triggers a finding with "credential" or "password" in the description |
+| `test_detects_private_key` | `-----BEGIN RSA PRIVATE KEY-----` header triggers a finding with "key" in the description |
+| `test_clean_code_no_findings` | Code that reads credentials from `os.environ` produces zero findings |
+| `test_detects_database_uri` | `postgres://user:pass@host/db` connection string triggers a finding with "connection string" or "credential" in the description |
+| `test_detects_github_token` | `ghp_…` GitHub Personal Access Token pattern triggers at least one finding |
+| `test_findings_have_gate_id` | Every finding from the AWS-key diff carries `gate == "secrets"` |
+| `test_findings_have_control_ids` | Every finding from the AWS-key diff uses a control ID within `{IA-5, IA-6, SC-12, SC-28}` |