AICARD

Auto-generated NIST AI RMF / EU AI Act Annex IV model & system cards

AI Security & Governance — securing LLMs, agents, and the MCP supply chain.

pip install cognis-aicard
aicard scan .            # → prioritized findings in seconds

🔎 Example output

Real, reproducible output from the tool — runs offline:

$ aicard-emit --version
aicard 0.3.8

$ aicard-emit --help
usage: aicard [-h] [--version] {check,card} ...

Auto-generate and lint NIST AI RMF / EU AI Act Annex IV model & system cards
from a JSON descriptor.

positional arguments:
  {check,card}
    check       evaluate a descriptor and report findings
    card        render a Markdown model card from a descriptor

options:
  -h, --help    show this help message and exit
  --version     show program's version number and exit

Example: aicard check demos/01-basic/system.json --format json

Blocks above are real aicard output — reproduce them from a clone.

Sample result format (illustrative values — run on your own data for real findings):

{"timestamp":1643723400,"data":{"indicator":"IP:192.168.1.100","description":"Suspicious network activity","severity":"high"},"findings":[{"id":123,"title":"Network Scan","description":"Network scan detected on 192.168.1.100","category":"network"},{"id":124,"title":"File Transfer","description":"File transfer detected from 192.168.1.100","category":"file_transfer"}]}

Related MCP server: asqav-mcp

Usage — step by step

aicard auto-generates and lints NIST AI RMF / EU AI Act Annex IV model & system cards from a JSON descriptor.

1. Install (Python 3.10+):

   pip install -e .            # or: pipx install aicard

2. Check a descriptor against the disclosure requirements (human-readable table):

   aicard check demos/01-basic/system.json

3. Render a Markdown model/system card from the same descriptor:

   aicard card system.json > MODEL_CARD.md

4. Read the output in the format your workflow speaks — table (default), json, sarif (SARIF 2.1.0 for code-scanning), or csv (GRC dashboards):

   aicard check system.json --format json  | jq '.findings'
   aicard check system.json --format sarif > aicard.sarif   # upload to GitHub code-scanning
   aicard check system.json --format csv   > findings.csv   # drop into a model-risk tracker
   aicard card  system.json --format json  | jq -r '.card_markdown'

5. Gate CI on compliance — check/card exit 1 when any blocking finding is present, 0 when compliant, 2 on input error:

   - run: pip install -e . && aicard check system.json   # non-zero fails the job

Worked demos

demos/ ships realistic descriptors in the real JSON input format, each with a SCENARIO.md (provenance, expected output, exact run command, how to act):

aicard check demos/14-insurance-pricing/auto_pricing.json --format csv

Contents

• Why aicard? · Features · Quick start · Example · Architecture · AI stack · How it compares · Integrations · Install anywhere · Related · Contributing

Why aicard?

Auto-generated NIST AI RMF / EU AI Act Annex IV model & system cards — without standing up heavyweight infrastructure.

aicard is single-purpose, scriptable, and self-hostable: point it at a target, get prioritized results in the format your workflow already speaks (table · JSON · SARIF), gate CI on it, and let agents drive it over MCP.

Features

• ✅ Load Descriptor

• ✅ Evaluate against 18 NIST AI RMF / EU AI Act Annex IV disclosure requirements

• ✅ Render Card (Markdown model/system card)

• ✅ Render Report Table

• ✅ Export findings as JSON · SARIF 2.1.0 · CSV

• ✅ Report To Dict

• ✅ 8 worked demos in demos/ (credit, fraud, medical, EdTech, ADAS, insurance, recsys, GenAI)

• ✅ Runs on Linux/macOS/Windows · Docker · devcontainer

• ✅ Ports in Python, JavaScript, Go, and Rust (ports/)

Quick start

pip install cognis-aicard
aicard --version
aicard scan .                       # scan current project
aicard scan . --format json         # machine-readable
aicard scan . --fail-on high        # CI gate (non-zero exit)

Example

$ aicard scan .
  [HIGH    ] AIC-001  example finding             (./src/app.py)
  [MEDIUM  ] AIC-002  another signal              (./config.yaml)

  2 findings · risk score 5 · 38ms

Architecture

flowchart LR
  IN[input] --> P[aicard<br/>analyze + score]
  P --> OUT[report]

Use it from any AI stack

aicard is interoperable with every popular way of using AI:

• MCP server — aicard mcp (Claude Desktop, Cursor, Cognis.Studio, uncensored-fleet)

• OpenAI-compatible / JSON — pipe aicard scan . --format json into any agent or LLM

• LangChain · CrewAI · AutoGen · LlamaIndex — wrap the CLI/JSON as a tool in one line

• CI / scripts — exit codes + SARIF for non-AI pipelines

How it compares

Integrations

Pipes into your stack: SARIF for code-scanning, JSON for anything, an MCP server (aicard mcp) for AI agents, and a webhook forwarder for SIEM/Slack/Jira. See docs/INTEGRATIONS.md.

Install — every way, every platform

pip install "git+https://github.com/cognis-digital/aicard.git"    # pip (works today)
pipx install "git+https://github.com/cognis-digital/aicard.git"   # isolated CLI
uv tool install "git+https://github.com/cognis-digital/aicard.git" # uv
pip install cognis-aicard                                          # PyPI (when published)
docker run --rm ghcr.io/cognis-digital/aicard:latest --help        # Docker
brew install cognis-digital/tap/aicard                             # Homebrew tap
curl -fsSL https://raw.githubusercontent.com/cognis-digital/aicard/main/install.sh | sh

Related Cognis tools

• aegis — AI Agent Permission & Access Auditor — surfaces the lethal trifecta of credentials + injection + reach

• promptmirror — Prompt-injection & indirect-injection scanner for any LLM context input

• ledgermind — Local LLM cost & token forensics proxy with anomaly detection

• adversa — LLM red-team harness — OWASP LLM Top 10 + MITRE ATLAS attack packs

• guardpost — Runtime agent firewall — PII redaction, rate limits, policy enforcement

• hallumark — LLM hallucination & grounding auditor for RAG systems

Explore the suite → 🗂️ all 170+ tools · ⭐ awesome-cognis · 🔗 cognis-sources · 🤖 uncensored-fleet · 🧠 engram

Contributing

PRs, new rules, and demo scenarios are welcome under the collaboration-pull model — see CONTRIBUTING.md and SECURITY.md.

⭐ If aicard saved you time, star it — it genuinely helps others find it.

Interoperability

{} composes with the 300+ tool Cognis suite — JSON in/out and a shared OpenAI-compatible /v1 backbone. See INTEROP.md for the suite map, composition patterns, and reference stacks.

License

Source-available under the Cognis Open Collaboration License (COCL) v1.0 — free for personal, internal-evaluation, research, and educational use; commercial / production use requires a license (licensing@cognis.digital). See LICENSE.