BC tools
Building LLM tools with Vovk.ts — `deriveTools()` (procedures → tools) returning a `StandardToolV0[]` array, standalone tools via the `standard-tool` package, `@operation.tool({ name, title, description, hidden, meta })` decorator, `x-tool` metadata, `ToModelOutput` formatters (`DEFAULT`, `MCP`, `RAW`), the merged `inputSchema`, the `meta` option for passing context to procedures, `withDefaults` for baking auth into third-party API tools, and the `pick` / `omit` pattern for selecting specific procedures. Use whenever the user asks to expose an API to an LLM, wire up tool calling, build an MCP server, "let Claude / GPT call this", "turn my controllers into tools", "function calling with Vercel AI SDK", "standalone tool", "hide a procedure from the LLM", "tool that wraps a third-party SDK", or any variation. Does NOT cover procedure authoring → `procedure` skill. Does NOT cover the OpenAPI spec beyond `@operation`/`x-tool` → `openapi` skill. Does NOT cover third-party OpenAPI mixin setup → `mixins` skill.
Promises to convert procedures into LLM tools via `deriveTools()` — generates `StandardToolV0[]` with `@operation.tool()` decorator support, `ToModelOutput` formatters (DEFAULT, MCP, RAW), and `pick`/`omit` patterns. The code implements basic schema transformation, but the description cuts off mid-sentence—claims more than files deliver.
Tests scored B: quality 79%, process 54%. No critical issues, but process coverage is weak. Runs on major platforms (Claude, GPT, Gemini, Cursor), though MCP and formatter integration quality remains untested—no model runs recorded. Works for straightforward tool calling in Claude; complex scenarios with custom logic may need fixes.
Building LLM tools with Vovk.ts — deriveTools() (procedures → tools) returning a StandardToolV0[] array, standalone tools via the standard-tool package…
As a process C 54/100 · Has gaps — weak spots: result and completion, inputs and preconditions, running it twice
How to improve
- The SKILL.md body is over 5,000 tokens: move reference detail into references/ and load it when needed.
- Your own cases (evals/evals.json, 4–6 real requests with expected answers): the full check would then run those instead of a model-drafted suite.
- A spec.yaml with trigger phrases and assertions — a behaviour contract for CI; `skilltest init` writes a template.
Guard findings · 0
✓ No critical or high findings
Files scanned: 1. Evidence is masked. Grey chips explain why severity was lowered.
Against the Agent Skills spec
- warning
body-longSKILL.md body ≈ 6023 tokens (recommended < 5000); move details to references/
Process rating: all ten parameters 54/100
- 0Result and completion. Does not say what the result is
- 0Inputs and preconditions. Does not say what the process needs to start
- 0Progress reporting. Says nothing while it works
- 30Running it twice. 13 mutating operations with no state check
- 50Failures and branches. 0 branches, has a failure section
- 60Tools and files. Uses tools (web) that frontmatter does not declare
- 70Execution cost. Instruction body is 6023 tokens
- 85Steps. 20 steps, 1 vague phrases
- 100When it triggers. States when to use and when not to
- 100Consistency. Name and required fields are in place
- low The response is described with custom markup (3 tags): a typed call is more reliable
Everything here is measured from the skill text rather than judged by a model, so the numbers are checkable. A parameter weighs more when it is a more common reason for the process to stall.
Quality signals
- +4Description does not say when NOT to use the skill (false activations)
- +3Description length 1018: 120–800 characters recommended
- +1No license
- +2Single-language instructions
- +5Description quotes 6 example trigger phrases
- +4Structure: 19 headings
- +3Step-by-step instructions: 20 items
- +3Output format is stated explicitly
- +4Has examples (19 code blocks)
Quality base 70; lint remarks subtract, signals add up to 100. Result: 79.