Workbench Integration

: Leveraging the dynamic execution layer for vault maintenance.

The workbench-integration (located in 00_Raw/workbench/) acts as the dynamic counterpart to the wiki's static knowledge. It provides a persistent, auditable environment for repo-aware investigations and "Executable Knowledge" verification.

1. The Proof-Round Workflow

The "Proof Round" is the primary mechanism for generating verifiable evidence from a codebase.

Action: Run node 00_Raw/workbench/dist/cli/main.js --host 00_Raw/workbench/dist/host/node-child-process-runner.js proof --repo <path>
Output: A JSON artifact that catalogs workspace state, entrypoints, and test health.
Leverage: These artifacts provide the ground truth for Literature Notes. When ingesting a new tool, run a proof round first to validate its claims before committing them to the wiki.

2. Thought API + Workbench

By combining the poshwiki-tools (the "Thought API") with Workbench's JavaScript REPL, agents can log their interactive discovery process in real-time.

Example Integration Script

// .workbenchrc.js
import { execSync } from 'child_process';

global.logThought = (content) => {
    const cmd = `pwsh -ExecutionPolicy Bypass -File 02_System/poshwiki-tools.ps1 -Content "${content}"`;
    execSync(cmd);
};

// Now in the Workbench REPL:
// > logThought("Discovered that vulture-mcp uses sqlx for SQLite.")

3. Leverage-Maximizing Strategies

A. Automatic Ingestion

Use Workbench to scan a repository and generate YANP-compliant drafts.

Run workbench eval "await repoAudit()" to get a structured JSON summary.
Use a script to map that JSON to a Fleeting Note.
Promote to a Permanent Note after human/librarian review.

B. Verification Loops

Treat wiki notes as "Type-Checked" artifacts.

A wiki note for a tool (like vulture-mcp) should include a Workbench Script to verify its health.
Run this script during vault maintenance (ps-vault-maintenance) to ensure the documentation hasn't drifted from the implementation.

C. Auditable Handoffs

When ending a session, include the Workbench audit-log.jsonl in your System Log. This provides the next agent with a "replay" of the discovery logic, reducing re-work and context drift.