---
title: "delegation-charter-diagnostic"
description: "Tests a LibreChat deployment against the Delegation Charter Specification v0.1.1: every agent against the eight pillars, the deployment against conformance Levels 1-3, with observe-only hooks that record what agents actu"
canonical: https://agentpluginsdirectory.com/plugins/delegation-charter-diagnostic
last-updated: 2026-10-10
---

# delegation-charter-diagnostic
Tests a LibreChat deployment against the Delegation Charter Specification v0.1.1: every agent against the eight pillars, the deployment against conformance Levels 1-3, with observe-only hooks that record what agents actually do.
- Slug: delegation-charter-diagnostic
- Publisher: Almma.AI
- Repository: https://github.com/almma-ai/librechat-charter-diagnostic
- Manifest: plugin.json
- Version: 0.1.1
- License: MIT
- Category (editorial): other
- Skills: 0
- MCP servers: 0
- Stars: 3
- Repository created: 2026-10-08
- Repository last pushed: 2026-10-09
- Publisher type: Organization
- Listing: https://agentpluginsdirectory.com/plugins/delegation-charter-diagnostic
- Schema: https://agent-plugins.org/schemas/1.0.0/plugin.schema.json

## What delegation-charter-diagnostic does, in the publisher's words

Test your LibreChat deployment against the Delegation Charter Specification: for every agent, is it written down what it will never do, when it stops, and who answers for it? And can your deployment hold agents to those terms?

It is an Agent Plugin you drop into your LibreChat fork, plus a one-command diagnosis. It reads; it never changes anything.

The sample above comes from the fictional demo database in test/fixtures, with simulated observer records and probe replies from a mock API. See the full sample report.

- Every agent against the eight pillars. Owner, escalation recipient, approval authority, never-do list, systems and permissions, inputs, escalation rules and execution constraints, plus the charter header and lifecycle rules. Each finding cites its evidence and says whether the term is enforced, only written in the instructions, or missing.

From the project README, punctuation lightly normalized. Full text: https://raw.githubusercontent.com/almma-ai/librechat-charter-diagnostic/HEAD/README.md