DTLaudit - Reference Manual
Descriptive reference for the local audit tool used across DTL projects.
Type: read-only auditLanguage: PythonCode version: v1.0-9Component Role
DTLaudit provides a control view over one project or a full suite of projects. It aggregates structure, documentation, Git status and optional GitHub information into a concise consistency report.
The tool observes only. It does not repair repositories, clean files or publish anything.
Functional Scope
Project audit
Describes one directory: Git presence, branch state, remote, tags, README files, documentation and significant files.
Suite audit
Compares direct subprojects under a root folder and highlights structural differences between them.
GitHub audit
When GitHub CLI is available, the report can include remote visibility, latest release, open pull requests and issues.
Internal Architecture
| Block | Description |
|---|---|
| Discovery | Identifies project candidates using local markers such as Git, Python, Windows scripts or README files. |
| Git collection | Reads current branch, remote, tags, local status and minimal history. |
| Documentation audit | Checks README files, user guides and reference manuals. |
| Rendering | Generates HTML by default, with optional JSON and text outputs. |
Observed Data
- Non-ignored files and folders.
- Generated directories such as
build,distand Python caches. - Large files above the configured threshold.
- Local Git state and available GitHub metadata.
- README, guide and reference documentation presence.
Output Model
The HTML report is the main output. It includes a global summary, a detailed block per project, notable observations and a file matrix that helps spot differences between projects.
The JSON output keeps a structure suitable for dashboards or automated processing. The text output is intended for quick reading.
Safety and Non-Mutation
DTLaudit is read-only. External commands are limited to reading Git and GitHub information. Audited files are not modified by the audit process.
Known Limits
- GitHub metadata depends on the installation and authentication state of
gh. - Project detection is intentionally simple and based on local markers.
- Generated artifacts tracked by Git can add noise to findings.