AB Data Construction Skill
build concept, process, and case-application supervision datasets from markdown books or long markdown documents. use when generating training data from many .md files or precomputed chunk files and when full chunk coverage, resumable batch processing, status tracking, validation, and coverage auditing are required. use for book-to-sft pipelines where every chunk must end in exactly one final status and where answer-only qa is not sufficient because the dataset should also teach grounded reasoning patterns and rule application.
As a process B 67/100 · Nearly there — weak spots: when it triggers, consistency, running it twice
How to improve
- 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: 10. Evidence is masked. Grey chips explain why severity was lowered.
Against the Agent Skills spec
- warning
name-formatname should be kebab-case (lowercase letters, digits, hyphens)
Process rating: all ten parameters 67/100
- 0Progress reporting. Says nothing while it works
- 20When it triggers. No condition that starts the skill
- 30Running it twice. 4 mutating operations with no state check
- 40Consistency. Frontmatter name (Data Construction Skill) differs from the folder (data-construction-skill)
- 60Result and completion. Output format stated, no completion criterion
- 70Inputs and preconditions. Inputs and preconditions are listed
- 70Failures and branches. 12 branches
- 70Execution cost. Instruction body is 4297 tokens
- 100Tools and files. No external tools needed
- 100Steps. 194 steps
- medium Safety rules and hard prohibitions inside a skill: they belong in the system prompt, here they protect nothing
- low 22 top-level sections: this looks like several domains in one skill
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
- +5Description has no quoted example phrases that should trigger the skill
- +4Description does not say when NOT to use the skill (false activations)
- -31 of 6 scripts are never mentioned in SKILL.md
- +1No license
- +2Single-language instructions
- +3Description length 533: enough signal without eating the budget
- +4Structure: 43 headings
- +3Step-by-step instructions: 194 items
- +3Output format is stated explicitly
- +4Has examples (10 code blocks)
Quality base 70; lint remarks subtract, signals add up to 100. Result: 79.