CB legitimate-interest
GDPR Legitimate Interest Assessment (LIA) — Guided assessment for Art. 6(1)(f) GDPR using the EDPB three-step test. Based on EDPB Guidelines 1/2024, EDPB Opinion 28/2024 (AI models), the OSS Case Digest on Legitimate Interest (McIntyre 2026), CNIL AI guidance (June 2025), ICO guidance, and key CJEU case law. Use when: (1) user needs to assess whether legitimate interest is appropriate, (2) user asks about Art. 6(1)(f), "Interessenabwägung", "berechtigtes Interesse", "balancing test", or "LIA", (3) user wants to document a legitimate interest assessment, (4) user mentions the three-step test, (5) user evaluates necessity or proportionality under Art. 6(1)(f), (6) user discusses direct marketing, fraud prevention, IT security, employee monitoring, AI training, web scraping, or credit checks in context of legal basis, (7) user handles a right to object (Art. 21), (8) user asks about children's data, profiling, or cross-border data sharing under legitimate interest, (9) user compares legitimate interest against other legal bases. Do NOT use for: consent assessments (Art. 6(1)(a)), contract performance (Art. 6(1)(b)), DPIA workflows, or pure ePrivacy/cookie compliance questions — though the skill does address the ePrivacy/GDPR overlap where relevant.
GDPR Legitimate Interest Assessment (LIA) — Guided assessment for Art.
As a process B 65/100 · Nearly there — weak spots: inputs and preconditions, consistency, execution cost
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: 11. Evidence is masked. Grey chips explain why severity was lowered.
Against the Agent Skills spec
- error
description-longdescription is 1266 chars, limit 1024 - warning
body-longSKILL.md body ≈ 8320 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
- 0Progress reporting. Says nothing while it works
- 40Consistency. Frontmatter name (legitimate-interest) differs from the folder (legitimate-interest-oliver-schmidt-prietz)
- 40Execution cost. Instruction body is 8320 tokens: crowds the task out of the window
- 50When it triggers. No condition that starts the skill
- 60Result and completion. Output format stated, no completion criterion
- 85Steps. 119 steps, 2 vague phrases
- 100Tools and files. No external tools needed
- 100Failures and branches. 1 branches, has a failure section
- 100Running it twice. Mutating operations check current state
- medium Safety rules and hard prohibitions inside a skill: they belong in the system prompt, here they protect nothing
- low 11 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
- +3Description length 1265: 120–800 characters recommended
- +1No license
- +2Single-language instructions
- +5Description quotes 3 example trigger phrases
- +4Description says when NOT to use the skill
- +4Structure: 28 headings
- +3Step-by-step instructions: 119 items
- +3Output format is stated explicitly
- +4Has examples (2 code blocks)
- +4Reference files are cited in the instructions (9 of 9)
Quality base 70; lint remarks subtract, signals add up to 100. Result: 62.