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

🔎 Example output

Real, reproducible output from the tool — runs offline:

$ checkpoint-ai-emit --version
CHECKPOINT-AI 0.1.0

$ checkpoint-ai-emit --help
usage: checkpoint-ai [-h] [--version] [--format {table,json,sarif,csv}]
                     {catalog,assess,ssp} ...

CHECKPOINT-AI: NIST AI RMF / EU AI Act / ISO 42001 self-assessment & SSP
generator.

positional arguments:
  {catalog,assess,ssp}
    catalog             list the cross-walked control catalog
    assess              score a self-assessment JSON file
    ssp                 generate an OSCAL-flavored SSP from an assessment

options:
  -h, --help            show this help message and exit
  --version             show program's version number and exit
  --format {table,json,sarif,csv}
                        output format (sarif/csv apply to the 'assess'
                        command)

$ checkpoint-ai-emit catalog
CHECKPOINT-AI control catalog (12 controls)
  GOV-1  (w5) GOVERN   AI governance policy & accountability owner
         nist_ai_rmf=GOVERN 1.1 eu_ai_act=Art.17 iso_42001=5.2
  GOV-2  (w4) GOVERN   Risk tolerance & escalation thresholds defined
         nist_ai_rmf=GOVERN 1.3 eu_ai_act=Art.9 iso_42001=6.1
  GOV-3  (w3) GOVERN   Workforce AI competency & training
         nist_ai_rmf=GOVERN 2.2 eu_ai_act=Art.4 iso_42001=7.2
  MAP-1  (w4) MAP      Intended purpose & context of use documented
         nist_ai_rmf=MAP 1.1 eu_ai_act=Art.11 iso_42001=8.1
  MAP-2  (w4) MAP      Foreseeable misuse & impacted populations identified
         nist_ai_rmf=MAP 3.1 eu_ai_act=Art.9 iso_42001=6.1.2
  MAP-3  (w5) MAP      Data provenance & lawful basis recorded
         nist_ai_rmf=MAP 2.3 eu_ai_act=Art.10 iso_42001=7.4
  MEA-1  (w4) MEASURE  Performance & accuracy metrics evaluated
         nist_ai_rmf=MEASURE 2.3 eu_ai_act=Art.15 iso_42001=9.1
  MEA-2  (w5) MEASURE  Bias / fairness testing across subgroups
         nist_ai_rmf=MEASURE 2.11 eu_ai_act=Art.10 iso_42001=9.1
  MEA-3  (w4) MEASURE  Adversarial robustness & security testing
         nist_ai_rmf=MEASURE 2.7 eu_ai_act=Art.15 iso_42001=8.3
  MAN-1  (w5) MANAGE   Human oversight & intervention controls
         nist_ai_rmf=MANAGE 1.1 eu_ai_act=Art.14 iso_42001=8.4
  MAN-2  (w4) MANAGE   Incident response & post-market monitoring
         nist_ai_rmf=MANAGE 4.1 eu_ai_act=Art.72 iso_42001=10.1
  MAN-3  (w4) MANAGE   Logging & traceability of system decisions
         nist_ai_rmf=MANAGE 2.2 eu_ai_act=Art.12 iso_42001=8.5

Blocks above are real checkpoint-ai output — reproduce them from a clone.

Usage — step by step

checkpoint-ai runs an AI-governance self-assessment cross-walked across NIST AI RMF, the EU AI Act, and ISO 42001, and can emit an OSCAL-flavored SSP.

1. Install:

   pip install -e .
   

2. List the cross-walked control catalog:

   checkpoint-ai catalog
   

3. Score a self-assessment JSON file:

   checkpoint-ai assess assessment.json
   

4. Generate an SSP (OSCAL-flavored System Security Plan) from the same assessment:

   checkpoint-ai ssp assessment.json > ssp.json
   

5. Export findings in the format your dashboard speaks. Each open control gap is a finding; assess can emit it as a table, JSON, SARIF 2.1.0, or CSV:

   checkpoint-ai --format sarif assess assessment.json > checkpoint.sarif.json   # code-scanning dashboards
   checkpoint-ai --format csv   assess assessment.json > gaps.csv                 # spreadsheets / GRC trackers
   

   The SARIF log carries one rule per control and one result per gap, each tagged with a security-severity and the NIST AI RMF / EU AI Act / ISO 42001 crosswalk.
6. Automate in CI — gate the build and publish findings on each governance change:

   checkpoint-ai assess assessment.json   # non-zero exit on an unaddressed weight-5 gap
   checkpoint-ai --format sarif assess assessment.json > checkpoint.sarif.json   # upload to code scanning
   

Demos — real-world scenarios

Each folder under demos/ is a self-contained scenario: a realistic assessment file in the tool's input format plus a SCENARIO.md describing where the data came from, what to expect, the exact run command, and how to act.

Demo	Scenario	EU tier	Outcome
01-pre-launch-gap-analysis	Support copilot six weeks from launch	limited	weight-5 gaps → CI fails
02-iso-42001-readiness	AIMS one control from a stage-1 audit	limited	single open gap
03-eu-ai-act-high-risk	Hiring system conformity prep (Annex III)	high	bias-testing gap blocks CE mark
04-prohibited-practice-stop	Citizen social-scoring proposal	unacceptable	perfect posture, still a hard stop
05-medical-device-high-risk	Retinal-screening triage that passes	high	clean — the target state
06-internal-analytics-minimal	Internal anomaly flagger	minimal	gaps, but exit 0 (right-sized)
07-greenfield-baseline	Day-zero assessment, nothing built	limited	0/100, full POA&M backlog
08-na-scoping	OCR digitizer scoping controls out	minimal	not_applicable handling
09-ci-sarif-gate	Credit adjudicator wired into CI	high	exit-code gate + SARIF upload
10-vendor-model-intake	Third-party SaaS due diligence	limited	gaps → vendor questionnaire

checkpoint-ai assess demos/09-ci-sarif-gate/self-assessment.json
checkpoint-ai --format sarif assess demos/09-ci-sarif-gate/self-assessment.json > checkpoint.sarif.json

Contents

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

Why checkpoint-ai?

NIST AI RMF / EU AI Act / ISO 42001 self-assessment & SSP generator — without standing up heavyweight infrastructure.

checkpoint-ai 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

• ✅ Cross-walked control catalog: NIST AI RMF · EU AI Act · ISO/IEC 42001
• ✅ Weighted posture scoring, maturity bands, and per-function breakdown
• ✅ EU AI Act risk-tier classification (minimal / limited / high / unacceptable)
• ✅ Gap analysis with prioritized remediation (OSCAL-flavored SSP + POA&M)
• ✅ Export findings as table · JSON · SARIF 2.1.0 · CSV
• ✅ CI gate: non-zero exit on an unaddressed high-weight gap
• ✅ 10 real-world demo scenarios in demos/
• ✅ Runs on Linux/macOS/Windows · Docker · devcontainer
• ✅ Ports in Python, JavaScript, Go, and Rust (ports/)

Quick start

pip install cognis-checkpoint-ai
checkpoint-ai --version
checkpoint-ai catalog                                            # list the cross-walked controls
checkpoint-ai assess assessment.json                             # score; non-zero exit on a high-weight gap
checkpoint-ai --format json  assess assessment.json              # machine-readable
checkpoint-ai --format sarif assess assessment.json              # SARIF 2.1.0 for code-scanning
checkpoint-ai ssp assessment.json > ssp.json                     # OSCAL-flavored SSP + POA&M

Example

$ checkpoint-ai assess demos/03-eu-ai-act-high-risk/self-assessment.json
CHECKPOINT-AI assessment: sentinel-resume-screener
  owner            : People Operations, Responsible AI Office
  EU AI Act tier   : high
  overall posture  : 72.8/100 (Defined)
  function scores  :
    GOVERN   : 91.2/100
    MANAGE   : 74.2/100
    MAP      : 76.2/100
    MEASURE  : 51.2/100
  framework cover  :
    nist_ai_rmf : 66.7%
    eu_ai_act   : 66.7%
    iso_42001   : 66.7%
  ...
  open gaps        : MAP-3, MEA-2, MEA-3, MAN-2
$ echo $?
2   # weight-5 gaps (MAP-3, MEA-2) remain — CI gate fails

Architecture

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

Use it from any AI stack

checkpoint-ai is interoperable with every popular way of using AI:

• MCP server — checkpoint-ai mcp (Claude Desktop, Cursor, Cognis.Studio, uncensored-fleet)
• OpenAI-compatible / JSON — pipe checkpoint-ai 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

	Cognis checkpoint-ai	usnistgov
Self-hostable, no account	✅	varies
Single command, zero config	✅	⚠️
JSON + SARIF for CI	✅	varies
MCP-native (AI agents)	✅	❌
Polyglot ports (JS/Go/Rust)	✅	❌
Open license	✅ COCL	varies

Built in the spirit of usnistgov/OSCAL, re-framed the Cognis way. Missing a credit? Open a PR.

Integrations

Pipes into your stack: SARIF for code-scanning, JSON for anything, an MCP server (checkpoint-ai 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/checkpoint-ai.git"    # pip (works today)
pipx install "git+https://github.com/cognis-digital/checkpoint-ai.git"   # isolated CLI
uv tool install "git+https://github.com/cognis-digital/checkpoint-ai.git" # uv
pip install cognis-checkpoint-ai                                          # PyPI (when published)
docker run --rm ghcr.io/cognis-digital/checkpoint-ai:latest --help        # Docker
brew install cognis-digital/tap/checkpoint-ai                             # Homebrew tap
curl -fsSL https://raw.githubusercontent.com/cognis-digital/checkpoint-ai/main/install.sh | sh

Linux	macOS	Windows	Docker	Cloud
scripts/setup-linux.sh	scripts/setup-macos.sh	scripts/setup-windows.ps1	docker run ghcr.io/cognis-digital/checkpoint-ai	DEPLOY.md (AWS/Azure/GCP/k8s)

Related Cognis tools

• cmmcmap — CMMC Level 2 practice mapper — stack-aware SSP skeleton generator
• fedramplens — FedRAMP boundary visualizer & OSCAL-format SSP/POAM generator
• sbirscout — SBIR/STTR topic discovery — DSIP + SBIR.gov + NIH digest with bid scoring
• gsafinder — GSA Schedule opportunity surveyor — SAM.gov + eBuy + FedConnect
• clearancepath — Personnel clearance hygiene tracker — SF-86, SEAD-3/4, training currency

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 checkpoint-ai 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 ([email protected]). See LICENSE.

---