A comprehensive whole-repository audit skill for AI agents to contrast stated system claims with the underlying reality of files, tests, and runtime evidence.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install project-truth-audit
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install project-truth-audit using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Project Truth Audit?
The Project Truth Audit skill empowers AI development agents to perform deep, objective evaluations of a repository's current state from the root directory. Instead of relying blindly on potentially outdated documentation, recent agent handoffs, or local module comments, this tool implements a rigid evaluation process that compares structural assertions against verifiable technical evidence.
By prioritizing active verification over passive assumptions, Openclaw Skills using this framework effectively mitigate the risks of tunnel vision and drift. The skill ensures that configuration requirements, entry points, deployment states, and code architectures are completely transparent and reconciled before any major refactoring or feature implementations occur.
Project Truth Audit Use Cases
- Whole-Repo Reality Check: Triggered when onboarding an AI agent to a codebase or when executing a high-level system overview.
- Contract Drift Identification: Used to cross-examine what a README or project specification document claims against what the codebase actually enforces.
- Stale Documentation and Half-Finished Audit: Helpful for scanning deep layers of code to isolate abandoned implementations, TODOs, or dead scripts.
- Pre-Implementation Guard: Deployed immediately before executing fixes to guarantee that dependencies, boundaries, and proof commands are accurately mapped.
How Project Truth Audit Works
- Resolve Root & Constraints: Read core project rules such as AGENTS.md, CLAUDE.md, or CONTEXT.md while honoring specific path or branch constraints.
- Construct the Truth Map: Inventory files using system commands (e.g., rg --files) to identify project claims, runtime enforcement points, and existing proof mechanisms like unit tests.
- Trace High-Risk Seams: Look beyond the immediate file context to review environment configurations, integration boundaries, authentication, and potential side effects.
- Evaluate and Classify Contracts: Map out major claims and rank their real-world statuses as verified, partial, stale, contradicted, or unknown.
- Generate Grounded Reports: Produce a severity-ordered findings report complete with proof commands and exact file paths prior to allowing any structural edits.
Project Truth Audit Setup
To leverage this skill within Openclaw Skills environments, include the skill definition file in your agent's directory layout.
Verify that essential system search tools like ripgrep are active within your execution context:
# Verify ripgrep availability for truth mapping
rg --version
# Execute a manual file discovery check from project root
rg --files
Ensure that configuration guidelines like AGENTS.md or CONTEXT.md are present at the root of the project to properly guide the scoping phase.
Project Truth Audit Data Schema & Taxonomy
The skill standardizes its findings into a clear, structured framework to prevent vague assessments. The output is organized into the following explicit data categories:
| Section | Type | Description |
|---|---|---|
| Scope | Metadata | Defines the requested root directory, active branch/ref, and explicit exclusions. |
| Read | File Paths | List of checked documentation files, build configs, entrypoints, and artifacts. |
| Proof | Execution Logs | Target commands or live probes run, complete with pass, fail, or blocked status indicators. |
| Findings | Severity List | Issues ranked by impact, tying claims directly to actual code realities and offering actionable fixes. |
| Truth Table | Structured Grid | Matrix aligning important claims with their actual enforcement, proofs, and conclusive status. |
| Unknowns | Exceptions | Gaps in verification with clear documentation detailing why confirmation was blocked. |
Project Truth Audit Advanced Features
- Multi-Layer Verification Matrix: Separates source-repo truth from live service endpoints, installable artifacts, and registry realities.
- Automated Proof Command Discovery: Requires the formulation of minimal validation commands to prove potential fixes work before making edits.
- Anti-Tunnel Vision Enforcement: Blocks agents from using nearby-file summaries as holistic whole-project context.
- Granular Status Tracking: Replaces generalized assessments with distinct evaluation states like partial, stale, contradicted, or unknown.
SKILL.md
Loading
Related Openclaw Skills
Alibaba Cloud Lindorm Vector Database Migration Assistant
An automated AI agent skill for seamless vector database migration from Milvus, Elasticsearch, Qdrant, or CSV to Alibaba Cloud Lindorm with built-in checkpoint resume.
Xiaohongshu Short Drama Trend Tracker (playlet-xhs-feed)
A smart feed integration for Openclaw Skills that tracks viral short drama trends on Xiaohongshu, offering dynamically clustered genre insights and professional HTML daily reports.
playlet-bili-feed
An automated tracking and intelligence tool that scans Bilibili for viral short drama content, sorts by engagement, clusters themes, and generates rich HTML reports.
PostFlow AI Content Hub
An all-in-one, multi-platform AI content management system that handles content generation, SEO scoring, platform adaptation, scheduling, and distribution.
UI Reviewer
An AI-powered frontend code reviewer that automates design system compliance, structural audits, responsiveness checks, and visual consistency for React and Vue applications.
Accessibility (A11y) Reviewer
An AI-powered web accessibility reviewer that automatically audits web pages and code structures against WCAG 2.2 AA compliance standards.
Featured*
