AC eval-integrity
Audit an LLM evaluation or benchmark repo for integrity and credibility practices. Use when asked to "audit my benchmark," "is my eval trustworthy," "check my leaderboard for contamination," "review this benchmark's methodology," or "what would a reviewer attack in my eval." Greps the target repo for evidence across seven dimensions (pre-registration, contamination, holdout hygiene, judge validity, statistical honesty, reproducibility, leaderboard exclusions) and emits a scored report with file:line evidence, severity, and concrete fixes.
Audit an LLM evaluation or benchmark repo for integrity and credibility practices.
As a process C 56/100 · Has gaps — weak spots: result and completion, when it triggers, inputs and preconditions
The same skill appears in 1 more place: ClawHub
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: 4. 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 56/100
- 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
- 20When it triggers. No condition that starts the skill
- 30Running it twice. 15 mutating operations with no state check
- 55Failures and branches. 1 branches
- 100Tools and files. No external tools needed
- 100Steps. 32 steps
- 100Consistency. Name and required fields are in place
- 100Execution cost. Instruction body is 2656 tokens
- low 10 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
- +4Description does not say when NOT to use the skill (false activations)
- +3Output format is not stated: the model decides each time
- +1No license
- +2Single-language instructions
- +5Description quotes 5 example trigger phrases
- +3Description length 544: enough signal without eating the budget
- +4Structure: 11 headings
- +3Step-by-step instructions: 32 items
- +4Has examples (2 code blocks)
Quality base 70; lint remarks subtract, signals add up to 100. Result: 89.