Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension


Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
8 changes: 8 additions & 0 deletions .github/workflows/release.yml
Original file line number Diff line number Diff line change
Expand Up @@ -39,6 +39,14 @@ jobs:
echo "skip=true" >> "$GITHUB_OUTPUT"
fi

- name: Require an empty [Unreleased] section
if: steps.exists.outputs.skip != 'true'
run: |
if [ "$(node -p "require('./dist/summary.json').unreleased_changes")" = "true" ]; then
echo "::error::CHANGELOG [Unreleased] still has notes; move them into the version section before releasing."
exit 1
fi

- name: Extract release notes from CHANGELOG
if: steps.exists.outputs.skip != 'true'
run: |
Expand Down
12 changes: 12 additions & 0 deletions CHANGELOG.md
Original file line number Diff line number Diff line change
Expand Up @@ -7,6 +7,18 @@ dataset version tracks `package.json` and `CITATION.cff` (kept in sync by

## [Unreleased]

## [0.6.0] - 2026-10-15

### Added
- `dist/summary.json` gained **`unreleased_changes`** (boolean). It is true
when `CHANGELOG.md`'s `[Unreleased]` section has content, meaning `main`
carries changes that no tagged release contains yet. The landing page shows
"plus unreleased changes" in that case. Before this, data merged between
releases appeared under the previous version number: after #18 the site
showed "0.5.0" for 100 records, while the v0.5.0 release holds 96.
- The release workflow refuses to tag while `[Unreleased]` still has notes.
CONTRIBUTING now asks every data PR to add an `[Unreleased]` entry.

### Data
- Triage of an external candidate brief (Rogue Agents Watch, 2026-10-15). Of
the candidates not already in the catalog, four passed re-validation against
Expand Down
2 changes: 1 addition & 1 deletion CITATION.cff
Original file line number Diff line number Diff line change
Expand Up @@ -10,7 +10,7 @@ authors:
website: "https://mlsecopshub.com"
repository-code: "https://github.com/MLSecOpsHub/agentic-attack-index"
url: "https://mlsecopshub.com"
version: 0.5.0
version: 0.6.0
date-released: 2026-10-15
license: CC-BY-SA-4.0
keywords:
Expand Down
3 changes: 2 additions & 1 deletion CONTRIBUTING.md
Original file line number Diff line number Diff line change
Expand Up @@ -49,6 +49,7 @@ Update the fields, add the new source(s), bump `last_updated`, and keep the `id`
## Commits and PRs

- Conventional Commits; data changes use the `data:` scope (e.g. `data: add <id>`).
- Every data PR adds a line under `## [Unreleased]` in `CHANGELOG.md`. That section drives `unreleased_changes` in `dist/summary.json`, so consumers can see when `main` is ahead of the latest tagged release.
- CI runs `npm test`; nothing merges red.
- Trust-critical paths (`data/`, `schema/`, `taxonomy/`, `scripts/`, `.github/`) have code owners (`.github/CODEOWNERS`). For this to actually gate merges, enable branch protection on `main` with **Require a pull request before merging**, **Require review from Code Owners**, and **Require status checks to pass** (select the `validate + build` check) — these are GitHub repo settings, not files in the repo.
- If you change `schema/`, update `README.md`, the taxonomy files, and `CITATION.cff` as needed, and explain the migration in the PR.
Expand All @@ -58,7 +59,7 @@ Update the fields, add the new source(s), bump `last_updated`, and keep the `id`
The dataset follows semantic versioning: added records or enum values are a minor bump; a removed or renamed field or enum value is a major bump with a migration note. To cut a release:

1. On a `release/vX.Y.Z` branch, set the same version in `package.json` (`npm version X.Y.Z --no-git-tag-version`, which also updates `package-lock.json`) and in `CITATION.cff`, along with `date-released`.
2. Move the `[Unreleased]` notes in `CHANGELOG.md` into a dated `[X.Y.Z]` section.
2. Move the `[Unreleased]` notes in `CHANGELOG.md` into a dated `[X.Y.Z]` section, leaving `[Unreleased]` empty. The release workflow refuses to tag otherwise.
3. Run `npm test` and `npm run linkcheck`, commit the rebuilt `dist/` as `chore(release): vX.Y.Z`, and open a PR.
4. After the merge, `.github/workflows/release.yml` creates the `vX.Y.Z` tag and a GitHub Release. The release notes come from the changelog section, and the `dist/` artifacts and schema are attached. The workflow fails if that section is missing.

Expand Down
2 changes: 1 addition & 1 deletion README.md
Original file line number Diff line number Diff line change
Expand Up @@ -38,7 +38,7 @@ npm run build # writes the dist/ artifacts below
| `dist/incidents.ndjson` | newline-delimited JSON (streaming) |
| `dist/incidents.csv` | flattened columns for spreadsheets (map points as `geo_points` = `role:basis:country`) |
| `dist/incidents/<id>.json` | one file per incident (stable per-record URL) |
| `dist/summary.json` | counts + rollups (category, severity, status, actor type, autonomy level, AI role, model family, year) + `archive_coverage` + `geo_coverage` (points by role and basis) + `evidence_split` (records that are both `confirmed` and `load-bearing`, overall and by year, plus a status × AI-role crosstab; retracted/superseded excluded) |
| `dist/summary.json` | `dataset_version` + `unreleased_changes` (true when `main` is ahead of that tagged release) + counts + rollups (category, severity, status, actor type, autonomy level, AI role, model family, year) + `archive_coverage` + `geo_coverage` (points by role and basis) + `evidence_split` (records that are both `confirmed` and `load-bearing`, overall and by year, plus a status × AI-role crosstab; retracted/superseded excluded) |
| `dist/stix/bundle.json` | STIX 2.1 bundle (reports + ATLAS/ATT&CK attack-patterns + CVE vulnerabilities) for TIP/MISP ingestion |
| `dist/index.html` | landing page (served via GitHub Pages) |

Expand Down
4 changes: 2 additions & 2 deletions dist/index.html
Original file line number Diff line number Diff line change
Expand Up @@ -3,7 +3,7 @@
<head>
<meta charset="utf-8">
<meta name="viewport" content="width=device-width, initial-scale=1">
<title>Agentic Attack Index — dataset 0.5.0</title>
<title>Agentic Attack Index — dataset 0.6.0</title>
<style>
:root{color-scheme:light dark}
body{font-family:system-ui,sans-serif;max-width:60rem;margin:2rem auto;padding:0 1rem;line-height:1.5}
Expand All @@ -15,7 +15,7 @@
</head>
<body>
<h1>Agentic Attack Index</h1>
<p>Open, source-linked dataset of real-world AI-agent cyberattacks. Dataset version <strong>0.5.0</strong>, 99 incidents.</p>
<p>Open, source-linked dataset of real-world AI-agent cyberattacks. Dataset version <strong>0.6.0</strong>, 99 incidents.</p>
<p class="artifacts">Artifacts:
<a href="./incidents.json">incidents.json</a>
<a href="./summary.json">summary.json</a>
Expand Down
3 changes: 2 additions & 1 deletion dist/summary.json
Original file line number Diff line number Diff line change
@@ -1,5 +1,6 @@
{
"dataset_version": "0.5.0",
"dataset_version": "0.6.0",
"unreleased_changes": false,
"schema": "https://raw.githubusercontent.com/MLSecOpsHub/agentic-attack-index/main/schema/incident.schema.json",
"total": 99,
"archive_coverage": {
Expand Down
4 changes: 2 additions & 2 deletions package-lock.json

Some generated files are not rendered by default. Learn more about how customized files appear on GitHub.

2 changes: 1 addition & 1 deletion package.json
Original file line number Diff line number Diff line change
@@ -1,6 +1,6 @@
{
"name": "agentic-attack-index",
"version": "0.5.0",
"version": "0.6.0",
"description": "Open, source-linked dataset of real-world cyberattacks executed or orchestrated by AI agents. Data layer for the Rogue Agent Watch tracker.",
"type": "module",
"private": true,
Expand Down
9 changes: 7 additions & 2 deletions scripts/build.mjs
Original file line number Diff line number Diff line change
Expand Up @@ -13,10 +13,13 @@ import {
sortKeysDeep,
} from './lib.mjs';
import { buildStixBundle } from './stix.mjs';
import { evidenceSplit } from './summary.mjs';
import { evidenceSplit, hasUnreleasedChanges } from './summary.mjs';

const pkg = JSON.parse(readFileSync(path.join(ROOT, 'package.json'), 'utf8'));
const schema = loadSchema();
const unreleasedChanges = hasUnreleasedChanges(
readFileSync(path.join(ROOT, 'CHANGELOG.md'), 'utf8')
);

const incidents = listIncidentFiles()
.map((file) => loadIncident(file))
Expand Down Expand Up @@ -58,6 +61,8 @@ const geoCoverage = {

const summary = {
dataset_version: pkg.version,
// main is ahead of the tagged dataset_version (CHANGELOG [Unreleased] non-empty).
unreleased_changes: unreleasedChanges,
schema: schema.$id,
total: incidents.length,
archive_coverage: {
Expand Down Expand Up @@ -143,7 +148,7 @@ code{font-size:.85em}
</head>
<body>
<h1>Agentic Attack Index</h1>
<p>Open, source-linked dataset of real-world AI-agent cyberattacks. Dataset version <strong>${esc(pkg.version)}</strong>, ${incidents.length} incidents.</p>
<p>Open, source-linked dataset of real-world AI-agent cyberattacks. Dataset version <strong>${esc(pkg.version)}</strong>${unreleasedChanges ? ' plus unreleased changes (see the CHANGELOG)' : ''}, ${incidents.length} incidents.</p>
<p class="artifacts">Artifacts:
<a href="./incidents.json">incidents.json</a>
<a href="./summary.json">summary.json</a>
Expand Down
9 changes: 9 additions & 0 deletions scripts/summary.mjs
Original file line number Diff line number Diff line change
Expand Up @@ -56,3 +56,12 @@ export function evidenceSplit(incidents) {
ids: clb.map((i) => i.id).sort(),
};
}

// True when CHANGELOG.md's [Unreleased] section has any content, i.e. main
// carries data or tooling changes that no tagged release contains yet. Lets
// consumers tell "dataset_version X" (a tagged release) apart from "main ahead
// of X" without a version bump on every data merge.
export function hasUnreleasedChanges(changelog) {
const m = changelog.match(/^## \[Unreleased\][^\n]*\n([\s\S]*?)(?=^## \[|(?![\s\S]))/m);
return Boolean(m && m[1].trim());
}
10 changes: 9 additions & 1 deletion test/summary.test.mjs
Original file line number Diff line number Diff line change
@@ -1,7 +1,7 @@
// Hermetic tests for the derived summary rollups in scripts/summary.mjs.
import { test } from 'node:test';
import assert from 'node:assert/strict';
import { evidenceSplit, activeRecords } from '../scripts/summary.mjs';
import { evidenceSplit, activeRecords, hasUnreleasedChanges } from '../scripts/summary.mjs';
import { listIncidentFiles, loadIncident } from '../scripts/lib.mjs';

const rec = (over = {}) => ({
Expand Down Expand Up @@ -69,3 +69,11 @@ test('evidence_split over the real dataset is internally consistent', () => {
assert.equal(yearSum, out.confirmed_load_bearing);
assert.ok(out.confirmed_load_bearing <= out.total);
});

test('hasUnreleasedChanges detects content under [Unreleased] only', () => {
const head = '# Changelog\n\nIntro.\n\n';
assert.equal(hasUnreleasedChanges(`${head}## [Unreleased]\n\n## [1.0.0] - 2026-01-01\n\n- x\n`), false);
assert.equal(hasUnreleasedChanges(`${head}## [Unreleased]\n\n### Data\n- added\n\n## [1.0.0] - 2026-01-01\n`), true);
assert.equal(hasUnreleasedChanges(`${head}## [Unreleased]\n- only section\n`), true);
assert.equal(hasUnreleasedChanges(`${head}## [1.0.0] - 2026-01-01\n- x\n`), false);
});
Loading