BB reverse-information-paradox
Diagnoses how consuming external AI forces you to reveal proprietary knowledge — your 'intelligence exhaust' (prompts, tool calls, corrections, evals) — to whoever controls the learning infrastructure, and prescribes a trust boundary that keeps your data, evals, memory, adapted weights, and learning loops inside your tenant while decoupling orchestration from any single provider. Activate when: user asks 'does using this AI vendor train their model on our data', 'are our corrections/evals improving a model our competitors also use', 'how do we adopt external AI without giving away our know-how', 'who owns the learning loop', 'are we locked into one AI provider', 'we're fine-tuning/RAG-ing on proprietary data with a third-party model', or describes an enterprise/SMB whose operational data flows into a vendor AI (coding assistant, support bot, vertical SaaS AI, fine-tuning partner). Do NOT activate when: the AI is fully local/self-hosted with no telemetry leaving your boundary and no shared learning; the question is about AI model *accuracy* or *prompt engineering* with no data-ownership/competitive-leakage dimension; or you are the AI *provider* designing a training pipeline (that is a different, seller-side problem). More: deciqai.com/c/reverse-information-paradox
Diagnoses how consuming external AI forces you to reveal proprietary knowledge — your 'intelligence exhaust' (prompts, tool calls, corrections, evals) — to…
As a process B 65/100 · Nearly there — weak spots: inputs and preconditions, failures and branches, running it twice
How to improve
- Shorten the description to 1024 characters.
- 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: 7. Evidence is masked. Grey chips explain why severity was lowered.
Against the Agent Skills spec
- error
description-longdescription is 1284 chars, limit 1024 - warning
body-longSKILL.md body ≈ 5255 tokens (recommended < 5000); move details to references/
Process rating: all ten parameters 65/100
- 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. 12 mutating operations with no state check
- 60Result and completion. Output format stated, no completion criterion
- 70Execution cost. Instruction body is 5255 tokens
- 85Steps. 46 steps, 1 vague phrases
- 100Tools and files. No external tools needed
- 100When it triggers. States when to use and when not to
- 100Consistency. Name and required fields are in place
- 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
- +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 1284: 120–800 characters recommended
- -41 reference files, but SKILL.md never points to them: the model will not open them
- +1No license
- +2Single-language instructions
- +4Structure: 12 headings
- +3Step-by-step instructions: 46 items
- +3Output format is stated explicitly
- +4Has examples (1 code blocks)
Quality base 70; lint remarks subtract, signals add up to 100. Result: 45.