Persistent note system with tags and search for AI agents. Reads free, writes require Veyra.
Server has 6 tools with complete schemas and descriptions. Naming follows verb_noun convention (list_, get_, search_, create_, update_, delete_). Descriptions are present but generic, most lack context on WHEN to use the tool or what happens on failure. Parameters are typed but lack constraints (enums, ranges, patterns). Critical issue: veyra_token is exposed as a parameter on write operations, violating secret-injection pattern. Output schemas are not documented. Error handling returns JSON but lacks recovery guidance.
Create a new note. Requires Veyra commit mode (Class A — €0.005).
Delete a note by ID. Requires Veyra commit mode (Class B — €0.02).
Retrieve a note by ID. FREE — no token required.
List all notes, optionally filtered by tag. FREE — no token required.
Search notes by query (matches title, content, tags). FREE — no token required.
Update an existing note's title or content. Requires Veyra commit mode (Class A — €0.005).
veyra_token exposed as tool parameter on create_note, update_note, delete_note. Credentials must never appear in parameters, they are logged in traces and prompt history. Use server-side secret injection via environment variables.
Output schemas not documented. LLMs cannot plan downstream calls or extract fields without knowing response structure. Document what list_notes, get_note, search_notes, create_note, update_note, delete_note return.
Descriptions lack context on WHEN to use each tool and what happens on error. E.g., create_note says 'Requires Veyra commit mode' but does not explain what happens if token is invalid or missing. Add recovery guidance.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | D | 59 | 2026-07-28+ | v2 |
Parameters lack constraints. 'limit' on list_notes has no min/max (unbounded). 'tags' on create_note is free-form string with no format guidance. Add enums, ranges, patterns, and length limits.
Error responses return JSON but lack actionable guidance. E.g., 'NoteNotFound' tells LLM nothing about what to do next. Include suggestions: 'Note not found. Try search_notes() to find the correct ID.'