AC doc-consistency
Check a long document against itself — dangling cross-references, numbering gaps and duplicates, TOC-vs-body mismatch, missing standard sections, terminology drift, leftover placeholders (TODO/TBD/XXX), stale years, and paragraphs lost in a merge. Reads .docx / .md / .txt and emits a located findings list plus a re-check command you can put in a contract. Runs fully offline with zero dependencies: no network, no upload, no model calls, so the document never leaves the machine. Use when the user asks to proofread, QA, sanity-check, or verify the internal consistency of a textbook, manuscript, bid or RFP response, manual, thesis, or contract set; or asks for a cross-reference audit, numbering audit, continuity check, or editorial consistency check; or says 长文档一致性 / 交叉引用 / 编号跳号 / 重号 / 术语不统一 / 合稿丢段 / 目录对不上 / 标书自检 / 稿件体检 / 定稿前检查 / 教材编号核对.
Check a long document against itself — dangling cross-references, numbering gaps and duplicates, TOC-vs-body mismatch, missing standard sections, terminology…
As a process C 52/100 · Has gaps — weak spots: result and completion, inputs and preconditions, failures and branches
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: 16. Evidence is masked. Grey chips explain why severity was lowered.
Against the Agent Skills spec
✓ No remarks against the Agent Skills spec
Process rating: all ten parameters 52/100
- 0Result and completion. Does not say what the result is
- 0Inputs and preconditions. Does not say what the process needs to start
- 0Failures and branches. Linear process with no failure handling
- 0Progress reporting. Says nothing while it works
- 60Tools and files. Uses tools (node) that frontmatter does not declare
- 70When it triggers. States when to use, but not when not to
- 100Steps. 15 steps
- 100Consistency. Name and required fields are in place
- 100Execution cost. Instruction body is 572 tokens
- 100Running it twice. No mutating operations
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)
- +3Description length 845: 120–800 characters recommended
- +3Output format is not stated: the model decides each time
- -5TODO / placeholder text left in the skill
- -31 of 2 scripts are never mentioned in SKILL.md
- +2Single-language instructions
- +4Structure: 8 headings
- +3Step-by-step instructions: 15 items
- +4Has examples (2 code blocks)
- +1License stated
Quality base 70; lint remarks subtract, signals add up to 100. Result: 74.