BC ensemble-rule-review
Design pattern for converting rule-following, checklist, or rubric skills into a fan-out map-reduce ensemble of parallel rigid sub-agents with corroboration-weighted merge; worker model tier and diversity are knobs matched to inference load and stakes, not fixed values. Apply when creating or refactoring a skill or agent that applies 10+ independent criteria in a single pass. Triggers on: 'review against a checklist', 'fan out', 'map reduce review', 'ensemble', 'split the rules', 'apply rubric', or any large ruleset being applied by one agent in one pass. NOT for tight single-pass transforms or rulesets under 5 criteria.
Design pattern for converting rule-following, checklist, or rubric skills into a fan-out map-reduce ensemble of parallel rigid sub-agents with…
As a process C 51/100 · Has gaps — weak spots: result and completion, inputs and preconditions, failures and branches
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: 17. Evidence is masked. Grey chips explain why severity was lowered.
Against the Agent Skills spec
- warning
body-longSKILL.md body ≈ 5802 tokens (recommended < 5000); move details to references/
Process rating: all ten parameters 51/100
- 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
- 30Running it twice. 10 mutating operations with no state check
- 60Tools and files. Uses tools (bash) that frontmatter does not declare
- 70Execution cost. Instruction body is 5802 tokens
- 100Steps. 40 steps
- 100When it triggers. States when to use and when not to
- 100Consistency. Name and required fields are in place
- low 12 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
- +3Output format is not stated: the model decides each time
- -31 of 5 scripts are never mentioned in SKILL.md
- +1No license
- +2Single-language instructions
- +4Description says when NOT to use the skill
- +3Description length 628: enough signal without eating the budget
- +4Structure: 15 headings
- +3Step-by-step instructions: 40 items
- +4Has examples (2 code blocks)
- +4Reference files are cited in the instructions (10 of 10)
Quality base 70; lint remarks subtract, signals add up to 100. Result: 79.