Trojaino v0.1.3

Blockhouse-branded HTML reports

• Restyles standalone HTML scan reports in the Blockhouse cream, charcoal, and orange visual system.
• Adds the Blockhouse header lockup, scan timestamp, and explicit Trojaino version metadata.
• Preserves bounded preflight, incomplete-scan, capability, and output-sanitization behavior.
• Documents Python 3.11+ and Tkinter requirements, plus the editable-install pip upgrade step.

No Trojaino package is published to PyPI. Install from this verified source release.

Trojaino v0.1.2 — Public Alpha

Trojaino v0.1.2 renames AI Shield and adds an optional desktop scan window plus optional, preview-first anonymous scan statistics.

Highlights:

• Optional Tkinter GUI for local file/folder scans with bounded Standard, Large, and Exhaustive presets
• HTML and JSON reports saved to a user-chosen location
• Optional anonymous aggregate statistics sharing; no source, report, path, filename, line-number, or evidence upload
• A one-time receipt and deletion token allow contributors to remove a submitted statistic
• New Trojaino name, report styling, and project branding

Trojaino remains a local deterministic install gate. It does not certify software as safe or replace human security review.

Install from a verified source checkout:

git clone https://github.com/BlockhouseSoftware/trojaino.git
cd trojaino
git checkout v0.1.2
python3 -m venv .venv
source .venv/bin/activate
python3 -m pip install -e .
trojaino gui

AI Shield v0.1.1 — Public Alpha

AI Shield v0.1.1 — Public Alpha

AI Shield v0.1.1 strengthens the local, deterministic pre-flight scan for AI-built and downloaded software. It remains source-only: no official PyPI package is published with this release.

What changed

• Hardened scanning of hostile repository inputs: bounded traversal, symlink and replacement-race defenses, explicit unreadable/invalid-file handling, bounded reports, and fail-closed behavior when coverage or rule execution is incomplete.
• Added metadata-only preflight estimates and finite scan-budget controls, including safe higher-budget recommendations.
• Added dependency-free Python-app heuristics for common high-risk execution, deserialization, configuration, route, and user-input patterns.
• Improved evidence sanitization and version-correct report metadata.

Verification

The exact tagged commit passed:

• 78 unit tests
• GitHub CI on Python 3.11, 3.12, 3.13, and 3.14
• Repository-boundary and CodeQL checks
• Release-profile self-scan: NO CRITICAL RISKS FOUND, complete, zero findings
• Clean wheel and source-distribution build
• Fresh wheel-install smoke tests: clean fixture exit 0; deliberately risky Python fixture exit 2
• Full-history Gitleaks scan: zero findings

Install from source

git clone --branch v0.1.1 https://github.com/BlockhouseSoftware/ai-shield.git
cd ai-shield
python3 -m aishield scan ./tests/fixtures/clean-project

Important limitations

AI Shield is an alpha deterministic scanner. It does not prove a project is safe, replace a full security audit, or detect every logic, dependency, obfuscation, generated-code, or runtime-only risk. Treat findings as evidence for review and DO NOT RUN as a stop-and-review signal.

Licensed under the GNU Affero General Public License v3.0 only (AGPL-3.0-only).

AI Shield v0.1.0 — Public Alpha

AI Shield v0.1.0 — public alpha

AI Shield is a local, deterministic trust scanner for AI-built and downloaded software. This first public release focuses on evidence-first checks before you run or ship a project.

Initial scope

• Node and TypeScript projects
• Package lifecycle and install scripts
• Docker and Compose configuration
• MCP servers and tool repositories
• Agent instruction files such as AGENTS.md, CLAUDE.md, and editor rules
• Obvious secret exposure and risky client-key patterns
• High-signal route, shell, evaluation, and filesystem heuristics

Install from source

git clone --branch v0.1.0 https://github.com/BlockhouseSoftware/ai-shield.git
cd ai-shield
python3 -m aishield scan ./tests/fixtures/clean-project

AI Shield v0.1.0 is distributed through this repository and GitHub release source archives only. No official PyPI package exists yet; do not install similarly named PyPI packages unless a future AI Shield release explicitly links to one.

Verification

• 47 unit tests passed
• CI passed on Python 3.11, 3.12, 3.13, and 3.14
• Repository-boundary check passed
• Release-profile self-scan returned NO CRITICAL RISKS FOUND with zero findings and zero capabilities
• Full public-history Gitleaks scan returned zero findings
• Fresh wheel build/install and clean/risky CLI smoke tests passed
• CodeQL passed for Python and GitHub Actions

Important limitations

AI Shield v0.1.0 is an alpha deterministic scanner. It does not prove that a project is safe, replace a full security audit, or detect every logic, dependency, obfuscation, generated-code, or runtime-only risk. Treat findings as evidence for review and DO NOT RUN as a stop-and-review signal.

Licensed under the GNU Affero General Public License v3.0 only (AGPL-3.0-only).