Cross-Cutting Tools

Plugin Admin

Administrative workspace for plugin configuration, diagnostics, and maintenance.

Download the CRE Skills Plugin

Latest release, portable bundle (signed). Review the SKILL.md files before installing into your agent.

dataNo personal data
What it does

A single admin workspace for plugin configuration, health diagnostics, workspace management, and usage reporting. Runs checks, routes maintenance tasks, and logs actions to an audit trail.

Why it matters

Practitioners who rely on a skill plugin need a reliable way to know it is working correctly, customize its defaults for their firm, and recover from broken or abandoned session workspaces. Without a dedicated admin surface those tasks scatter across config files, directory browsing, and trial-and-error.

How it's done today

A user manually inspects the config JSON, browses the workspaces directory, re-runs a skill to see if it fires correctly, and tracks nothing. Diagnostics are ad hoc and rarely leave any record of what was changed or why.

When to use it

Reach for it

Activate when you need to review or change plugin settings, run a health check after installation or update, manage saved workspaces, check skill inventory, or submit feedback and problem reports.

Not the right tool

Not for running CRE analysis workflows. Once the plugin is healthy and configured, route to the relevant workflow skill such as deal-quick-screen or lease-abstract-extractor rather than staying in the admin workspace.

Example use case

After upgrading the plugin, an analyst notices the router is not resolving deal workspace keywords. They invoke Plugin Admin, run catalog validation, find two SKILL.md files with missing frontmatter fields, and get a prioritized list of files to fix before re-routing to deal-quick-screen.

Compatible agents

Agent personas that pair well with this skill

Limitations

Diagnostic checks read the local plugin directory and config file. They cannot verify remote API connectivity or catch runtime errors that only surface when a skill is executed against live data. A human must review and act on any flagged findings.