owasp-guard-cli 0.2.1

OWASP Guard CLI: secure code inspector

OWASP Guard CLI

OWASP Guard is a prompt-engineered secure code inspection CLI for identifying OWASP Top 10 (2021) issues in source code.

It scans a repository or single file, performs chunked LLM analysis with a verification pass, then generates:

• report.json for pipelines and automation
• report.md for human review and audit

Key Features

• OWASP-focused findings with confidence scoring
• Chunk-based analysis with route/function-aware heuristics
• Verification pass to reduce false positives
• Deduplication of repeated findings
• Progress-rich terminal UI
• Broad multi-language and config/IaC file support

Requirements

• Python 3.11+
• Groq API key

Installation

From PyPI:

pip install owasp-guard-cli

From source:

pip install -r requirements.txt
pip install -e .

Quick Start

1. Configure API key:

owasp-guard init

2. Run scan:

owasp-guard scan ./project --output-dir outputs

3. Review outputs:

• outputs/report.json
• outputs/report.md

CLI Usage

Main help:

owasp-guard --help
owasp-guard help

Topic help:

owasp-guard help init
owasp-guard help scan
owasp-guard help reports
owasp-guard help errors

Alias entrypoint:

owasp help
owasp help scan

Common Commands

Initialize with provided key:

owasp-guard init --api-key gsk_xxx

Scan repo:

owasp-guard scan "D:\path\to\repo" --output-dir outputs

Scan single file:

owasp-guard scan "D:\path\to\file.py" --output-dir outputs

Fixed-scope evaluation:

owasp-guard scan "D:\path\to\repo" --max-files 10 --output-dir outputs

Custom model:

owasp-guard scan ./project --model llama-3.1-8b-instant

Report Schema Summary

report.json includes:

• tool metadata and report version
• methodology metadata
• scope metrics (total_files, total_chunks, total_findings)
• summary metrics (risk_score_10, severity and OWASP distributions)
• full findings list

report.md includes:

• executive summary and risk score
• OWASP and file impact distribution
• findings index and detailed evidence/fix sections
• priority action plan and remediation roadmap

Exit Codes

• 0: scan completed with no findings
• 1: scan completed with findings
• 2: invalid CLI usage
• 3: runtime error

Error Types

• Configuration Error: missing/invalid API key or unreadable config
• Input Error: bad path, unsupported file type, invalid --max-files, empty scope
• API Error: model/provider failures
• Report Error: output write/serialization failures
• Scan Error: chunking or scan pipeline failures

Repository Docs

• Project prompt evolution: PROMPT_LOG.md
• Tool benchmark worksheet: comparison.md
• Sample outputs: report.md, testotpt/report.md, juiceotpt/report.md