eld-record-memory-collector

Collect and verify project information by scanning your codebase structure, reading key files, and using multi-model validation to record verifiable claims about architecture, dependencies, config, and documentation. The Skill implements a three-loop workflow of collect, Codex-based verification, and hash-based validation to ensure accuracy.

Multi-model verification cross-checks collected project facts against multiple validation sources—Codex and hash-based checks—to produce provenance-enhanced claims. This approach reduces errors when indexing a project or building a knowledge base by confirming data integrity across independent verification loops.

Yes. During codebase onboarding, the Skill automatically scans structure, reads key files, and produces verified memory claims about project layout, dependencies, and configuration. This accelerates knowledge transfer by capturing and validating facts upfront rather than manual discovery.

The Skill outputs verifiable memory claims directly to pce-memory using observe, upsert, and provenance-enhanced APIs. This creates a persistent, auditable record of project facts that tracks how and when claims were validated.

The Skill scans project structure and key files across typical software projects to extract facts about architecture, dependencies, configuration, and documentation. Specific file formats and scanning scope depend on your project layout; the Skill validates discovered claims against source files.