Nodex is a Rust CLI tool for graph-based document management. Tool definitions are inferred from command structure rather than explicitly registered with schemas. Most tools lack visible input parameter schemas in the provided source. Descriptions are present but generic (10-50 chars for many tools). No output schemas documented. Error handling is not visible in the sample code. The tool set is well-organized (20 tools, clear responsibilities) but lacks the structured metadata required for LLM-driven agent integration.
Parse all in-scope docs and build the graph
Run validation rules
Structural diff between two git refs (added/removed nodes & edges, status transitions, field changes)
Emit the authoritative CLI invocation grammar (leaf paths, positional arity, payload-schema keys)
Emit the resolved document-locating surface (scope, output, parser, identity, initial status)
Emit the error-code and exit-code vocabularies (closed sets, for codegen)
Emit closed-enum manifest (kinds, statuses, per-field enums)
Emit the JSON Schema of every CLI envelope shape (for codegen)
No visible input schemas for 14 of 20 tools. Tools like 'init', 'status', 'query', 'lifecycle', 'report', 'migrate', 'rename', 'retarget', 'scaffold', and most export subcommands have no documented parameter types or constraints.
Tool definitions are inferred from CLI command structure (clap Subcommand enum) rather than explicitly registered with MCP tool metadata. No evidence of tool registration with names, descriptions, and schemas in a structured format.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 35 | 2026-07-28+ | v2 |
Emit the active-rules manifest (built-in + config-driven rules)
Emit the project's frontmatter JSON Schema (draft 2020-12)
What depends on what a diff changed: removed/modified nodes paired with their transitive dependents
Create a nodex.toml in current directory
Manage document lifecycle
Inject missing frontmatter into legacy docs
Search and explore the graph
Move file and update references
Generate reports
Repoint references from one node id to another (e.g. after a supersession)
Create a new document node with valid frontmatter
Report the graph snapshot's state (absent / unreadable / schema_mismatch / outdated / current)
Descriptions are generic and under 50 characters for most tools (e.g., 'Report the graph snapshot's state', 'Search and explore the graph'). These lack context for LLM selection and do not explain WHEN to use each tool vs. similar ones.
No output schemas documented. LLMs cannot plan downstream tool calls or extract required fields without knowing what each tool returns. E.g., does 'diff' return node IDs, edge changes, or both? Does 'query' return paginated results?
No error handling guidance visible. Error responses should tell LLMs what to do next (retry, ask user, or fail). No evidence of categorized errors (retryable vs. user-fixable vs. fatal) or actionable error messages in the sample code.