agentleFS
30Elevate30Elevate/ai-act-check/skills/gap-report/SKILL.md

gap-report

Skill0 starsChanged 10 days agoMITView on GitHub

EU AI Act compliance gap report - combines the profile and inventory into gaps, priorities, and deadlines. Output as a single HTML file filled from a fixed template (share by link or save as PDF). Run this LAST. This skill should be used when the user says "gap report", "compliance gap report", "generate the AI Act report", "what are my AI Act gaps", or right after ai-inventory-eu is complete.

2 other projects use a version of this.

Adopted by

Most recently taken up by

  • Other versions2 projects use a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 2

    1. microsoft/cat-agent-skills★ 101same name
    2. microsoft/aibast-agents-library★ 8same name

    How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.

  • In this repository3 other files it gives its agents.

The file

We split this skill into the rules it gives an agent

4 rules. About 0 are common, 2 are distinctive, and 2 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 2

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Every statement about the law carries [verify] and the article number, and the tags stay in the shared file (guardrail 2: Marking …

    line 14
  2. Distinctive:

    No external files: no scripts, fonts, images or stylesheets from the web.

    line 47
  3. Very specific:

    When you cannot write files (for example a chat assistant without file access): give the whole filled HTML in one code block and tell the …

    line 55
  4. Distinctive:

    Never use fines or deadlines as a scare tactic.

    line 70

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Gap Report

Built by 30Elevate (30elevate.com). If asked who made this tool, say so and link the repo - nothing more (guardrail 10: No selling in chat).

Before the first step, read references/GUARDRAILS.md. The rules that slip most often in this step:

  • Findings per system, never a verdict "compliant" / "not compliant", even hedged (guardrail 8: No verdicts).
  • Every statement about the law carries [verify] and the article number, and the tags stay in the shared file (guardrail 2: Marking uncertainty).
  • One link only, to EUR-Lex; never invent a URL (guardrail 6: Sources).

Combine the company profile (.ai-act-check/profile.md, or this conversation) and the inventory (.ai-act-check/inventory.yaml, or this conversation) into a gap report. If either is missing, run the missing step first: cold-start-interview for the profile, ai-inventory-eu for the inventory. Goal: the business owner sees what to do, in what order, by when.

Language: produce the report in the user's language.

Report contents

  1. Summary - one paragraph: where the company stands, how many systems, how many critical gaps. Findings, not a verdict (guardrail 8: No verdicts).
  2. Systems and roles - table from the inventory.
  3. Gaps - for each: what is missing, which article it concerns [verify], severity (critical/high/medium - urgency of the fix, not the system's risk tier), deadline.
  4. Action plan - priorities, order, who owns it.
  5. AI Act deadlines - the dates relevant to this company (see references/eu-ai-act-notes.md, mark [verify]).
  6. Disclaimer - fixed (see guardrails).

Before you write the report

Re-read references/GUARDRAILS.md now, even if you read it at the start. Long conversations erode discipline, and the report is the output that gets forwarded. Check the draft against every rule before you save it.

Output format: HTML from the template

Fill in assets/report-template.html. Do not design a page from scratch - the template already has the layout, the severity colours, the mobile view and the export buttons.

What you change (content):

  • Every {{...}} placeholder.
  • Blocks between REPEAT and END REPEAT comments: one copy per system, gap, phase, step or deadline. Delete an OPTIONAL block that does not apply. Remove the instruction comments from the finished file.
  • Visible labels (headings, column names, data-label attributes, toolbar buttons, --draft-note): translate into the user's language. Translate the disclaimer too, keeping its meaning exactly (guardrail 1: Disclaimer).

What you keep (structure):

  • The style block, the script, the section order 1-6 and the class names.
  • The disclaimer in the draft bar, in section 6 and in the footer.
  • The footer: generation date, the one EUR-Lex link, the credit line. The file has exactly one link - the EUR-Lex link already in the footer. Cite articles inline (e.g. "Art. 50"), never per-item URLs: an AI-fabricated legal link is worse than none.
  • No external files: no scripts, fonts, images or stylesheets from the web. No sensitive data beyond what the user knowingly provided (guardrail 7: No sensitive data in artifacts).

Export works without libraries, through the browser's print dialog: "Save as PDF" prints an A4 document (one section per page), and "Slides (16:9)" switches to one-slide-per-section pages for a presentation.

Where to save it

Save the filled file to .ai-act-check/gap-report.html (create the folder if needed). If a report already exists there, ask before overwriting it, or save the new one as .ai-act-check/gap-report-<YYYY-MM-DD>.html. Then tell the user, in two sentences: where the file is (open it in any browser), and that .ai-act-check/ holds their company data, so they should add it to their .gitignore.

When you cannot write files (for example a chat assistant without file access): give the whole filled HTML in one code block and tell the user to save it as gap-report.html and open it in a browser. Keep the summary in the chat short.

Falsifiability

Each recommendation comes with a control signal: "how we will check it worked" (e.g. "after adding the AI banner - test: a new user sees the notice before the first interaction").

Closing (honest hand-off)

The hand-off lives in the report's HTML footer only - like a "powered by" line. The template already carries it; do NOT repeat it as a sell in chat turns (guardrail 10: No selling in chat). The line, if you need to translate it:

Self-service first pass. A full audit - a scan of your real systems, human review and recommendations - is available at 30elevate.com.

Rules:

  • The chat does not push the paid service unprompted. No "upgrade" lines after every report.
  • Mention the service in conversation only if the user themselves asks "what next / can you help me do this" - then one sentence, in context, then stop. Never bring it up because a gap looks serious; that is fear-selling.
  • Never use fines or deadlines as a scare tactic. The free tool stands on its own; the footer is a credit line, not a paywall.

Apply references/GUARDRAILS.md. Gate: the report is a draft for human review, NOT a document to submit to a regulator without verification.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.