feat(security): schedule the advisory early-warning scan by Dongni-Yang · Pull Request #7398 · NVIDIA/NemoClaw

Dongni-Yang · 2026-07-22T22:27:10Z

Summary

The operational slice split out of #7354 per review: the scheduled advisory early-warning workflow and the proposed operating-policy documentation. Draft until product/security owners record sign-off on #7338 — merging this PR closes #7338 (it satisfies the issue's last unchecked acceptance criterion; every other criterion shipped in the merged #7354).

Why this is separate

#7354 review (correctly) held that a scheduled workflow with issues: write plus its operating policy defines a supported surface, which #7338's unchecked acceptance criterion reserves for product/security owners: supported historical-image scope, rescan ownership, alert destination, and response expectations. #7354 (now merged) carries only the non-operational pieces (correlation module, local scan CLI, per-audit provenance sidecars). This PR carries the operational pieces and stays draft until the owners' decision lands on #7338.

Changes

.github/workflows/advisory-early-warning.yaml: six-hourly + manual dispatch; paginated fetch of reviewed/unreviewed/malware advisories filtered by affects batches of the reviewed inventory; correlation via the scan CLI; signals reported into one rolling security-labeled issue (marker- and Actions-bot-author-matched, control-character-sanitized, deduped on advisory id + package). Non-blocking by design; permissions: contents: read, issues: write. Now also reconciles signals with NVD (folded in from feat(security): add the NVD reconciliation library and the #7276 post-mortem #7370 per the restructure): up to 20 CVE ids per run, 7 s spacing for the unauthenticated rate limit, cap logged when exceeded; NVD outages degrade to an NVD: unavailable annotation and never fail the run; annotations extend rolling-issue line text only, the dedupe key is unchanged.
docs/security/advisory-early-warning.md: adds the "Scheduled operation" section (including the NVD reconciliation cadence/degradation wording) and the "Proposed policy defaults (pending maintainer confirmation)" section. Wording flips from proposed to accepted once Detect public advisories before reviewed ecosystem mappings reach scanners #7338 sign-off is recorded.

Rebased onto main after #7354 merged; the NVD workflow step was folded in from #7370 when that PR was narrowed to its non-operational slice — keeping all operational surface in this one sign-off-gated PR. With this fold-in, the merged workflow satisfies #7338's desired behavior of using NVD as a supplementary reconciliation source.

Merge checklist

feat(security): advisory early-warning correlation and audit provenance records #7354 merged (this PR was stacked on its branch; now rebased onto main)
feat(security): add the NVD reconciliation library and the #7276 post-mortem #7370 merged first — the workflow invokes the scan CLI's --nvd-records flag, which feat(security): add the NVD reconciliation library and the #7276 post-mortem #7370 introduces; merging this PR before feat(security): add the NVD reconciliation library and the #7276 post-mortem #7370 makes the six-hourly run fail on an unknown flag
Product/security owners define scope, rescan ownership, alert destination, and response expectations on Detect public advisories before reviewed ecosystem mappings reach scanners #7338 (proposed defaults are in the doc here)
Doc wording updated from proposed to accepted per that decision

Closes #7338

Signed-off-by: Dongni-Yang dongniy@nvidia.com

🤖 Generated with Claude Code

Summary by CodeRabbit

New Features
- Added scheduled security advisory monitoring every six hours, with manual runs available.
- Fetches npm security advisories (reviewed, unreviewed, malware), correlates them with the reviewed inventory, and publishes new alerts as deduplicated signals in a rolling, security-labeled issue.
- Enriches signals with NVD CVE annotations when available (gracefully marks enrichment as unavailable on rate limits/outages).
Documentation
- Expanded documentation for the scheduled operation, rolling issue lifecycle/append-only behavior, audit provenance, and proposed policy defaults requiring recorded sign-off.

Public advisories can affect maintained packages weeks before they reach the reviewed corpus that npm audit enforces (#7276 timeline). Add an early-warning path and per-audit provenance so those windows become visible and provable from retained artifacts: - scripts/lib/advisory-early-warning.mts: dependency-free correlation of GitHub Security Advisory records (reviewed, unreviewed, and malware) against the reviewed-graph package inventories; exact npm name and parseable-range matches produce investigate signals, ambiguous or CPE-style matches stay informational and never block. - .github/workflows/advisory-early-warning.yaml: six-hourly paginated scan of all three advisory types, correlated against the inventory, reporting into one rolling security-labeled issue (marker- and author-verified, deduplicated by advisory id plus package). - scripts/audit-reviewed-npm-graph.mts and the wechat runtime audit action now write a *.provenance.json sidecar next to each raw report: scanner and npm/node versions, configured registry and derived bulk advisory endpoint, run start/end times, extracted advisory ids, and a failure marker whenever the audit attempt itself failed - including npm's parseable-error-JSON failure mode - so a failed attempt can never be mistaken for a comparable non-detection. - docs/security/advisory-early-warning.md: how the path works plus proposed defaults (scope, ownership, alert destination, SLA) pending maintainer confirmation. Refs #7338 Signed-off-by: Dongni-Yang <dongniy@nvidia.com> Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

CodeRabbit: every source file carries the SPDX header; match the sibling docs/security file comment syntax. Refs #7338 Signed-off-by: Dongni-Yang <dongniy@nvidia.com> Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

…ovenance

…provenance

Narrow #7354 to its non-operational pieces per review: the correlation module, the local scan CLI, and the per-audit provenance sidecars. The six-hourly workflow (issues: write) and the proposed policy defaults move to a follow-up PR that stays draft until product/security owners sign off on #7338 scope, ownership, alert-destination, and response questions. Refs #7338 Signed-off-by: Dongni-Yang <dongniy@nvidia.com> Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

Reinstates the operational slice split out of #7354: the six-hourly workflow (permissions: contents read, issues write) that fetches reviewed, unreviewed, and malware advisories naming reviewed-inventory packages, correlates them with the scan CLI, and routes signals into one rolling security-labeled issue; plus the proposed policy defaults for #7338 open policy questions. Stays draft until product/security owners record sign-off on #7338. Refs #7338 Signed-off-by: Dongni-Yang <dongniy@nvidia.com> Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

copy-pr-bot · 2026-07-22T22:27:12Z

Auto-sync is disabled for draft pull requests in this repository. Workflows must be run manually.

Contributors can view more details about this message here.

coderabbitai · 2026-07-22T22:27:15Z

📝 Walkthrough

Walkthrough

The pull request adds a scheduled GitHub Actions workflow that collects and deduplicates npm advisories, correlates them into signals, enriches CVE-bearing signals with bounded NVD data, and maintains a rolling security issue. Documentation describes the operation and proposed policy defaults.

Changes

Advisory early-warning automation

Layer / File(s)	Summary
Schedule and advisory collection `.github/workflows/advisory-early-warning.yaml`	Adds scheduled and manual execution, inventory batching, paginated retrieval across reviewed, unreviewed, and malware advisories, and `ghsa_id` deduplication.
Signal correlation and NVD enrichment `.github/workflows/advisory-early-warning.yaml`	Generates `signals.json`, reconciles up to 20 CVEs with rate-limited NVD requests, records unavailable or skipped lookups, and incorporates annotations without changing signal deduplication keys.
Rolling security issue routing `.github/workflows/advisory-early-warning.yaml`	Creates or updates a marked open `security` issue with newly observed deduplicated signals, while exiting successfully without issue changes when no signals exist.
Scheduled operation and policy documentation `docs/security/advisory-early-warning.md`	Documents workflow cadence, advisory routing, NVD fallback behavior, issue reuse and closure rules, sign-off requirements, and proposed response policies.

Estimated code review effort: 4 (Complex) | ~45 minutes

Sequence Diagram(s)

sequenceDiagram
  participant GitHubActions
  participant AdvisoryScanner
  participant GitHubAdvisoriesAPI
  participant NVD
  participant GitHubIssuesAPI

  GitHubActions->>AdvisoryScanner: List inventory packages
  AdvisoryScanner->>GitHubAdvisoriesAPI: Fetch paginated advisories
  GitHubAdvisoriesAPI-->>AdvisoryScanner: Return advisory records
  AdvisoryScanner-->>GitHubActions: Write advisories.json and signals.json
  GitHubActions->>NVD: Reconcile capped CVE set
  NVD-->>GitHubActions: Return NVD records or unavailable status
  GitHubActions->>AdvisoryScanner: Incorporate NVD annotations
  AdvisoryScanner-->>GitHubActions: Write final signals.json
  GitHubActions->>GitHubIssuesAPI: Find, create, or update rolling security issue

Possibly related PRs

NVIDIA/NemoClaw#7370: Describes the related NVD reconciliation and signal annotation flow.

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 inconclusive)

Check name	Status	Explanation	Resolution
Linked Issues check	❓ Inconclusive	Core early-warning workflow is present, but the summary doesn't confirm the fixture, provenance, and rescan requirements from `#7338`.	Provide evidence or implementation for the pre-review fixture, provenance capture, and rescan-trigger behavior required by `#7338`.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Out of Scope Changes check	✅ Passed	Changes stay within the security workflow and documentation scope, with no unrelated feature work evident.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title clearly summarizes the main change: adding a scheduled advisory early-warning scan.

✨ Finishing Touches

📝 Generate docstrings

Create stacked PR
Commit on current branch

🧪 Generate unit tests (beta)

Create PR with unit tests
Commit unit tests in branch dongniy/7338-early-warning-workflow

_{Comment @coderabbitai help to get the list of available commands.}

github-code-quality · 2026-07-22T22:27:54Z

Code Coverage Overview

Languages: TypeScript

TypeScript / code-coverage/plugin

The overall coverage in commit 045b81c in the dongniy/7338-early-w... branch remains at 96%, unchanged from commit 04e6dfa in the main branch.

TypeScript / code-coverage/cli

The overall coverage in commit 045b81c in the dongniy/7338-early-w... branch remains at 80%, unchanged from commit ccf3dd4 in the main branch.

Show a code coverage summary of the most impacted files.

File	`main` ccf3dd4	`dongniy/7338-early-w...` 045b81c	+/-
`src/lib/inferen...lama/process.ts`	100%	50%	-50%
`src/lib/inferen...er-lifecycle.ts`	71%	65%	-6%
`src/lib/state/m...-acquisition.ts`	89%	84%	-5%
`src/lib/shields/index.ts`	72%	71%	-1%
`src/lib/state/sandbox.ts`	76%	76%	0%
`src/lib/state/o...d-extensions.ts`	98%	98%	0%
`src/lib/platform.ts`	84%	89%	+5%
`src/lib/onboard...ndbox-create.ts`	83%	91%	+8%
`src/lib/onboard...-create-plan.ts`	75%	88%	+13%
`src/lib/onboard...ndbox-create.ts`	33%	83%	+50%

_{Updated July 24, 2026 07:32 UTC}

github-actions · 2026-07-22T22:35:01Z

PR Review Advisor — Informational

Advisor assessment: Informational / high confidence
Next action: Review the warnings below.
Findings: 0 blockers · 1 warning · 0 suggestions
Status: Canonical ledger: 0 blocker(s), 1 warning(s), 0 suggestion(s).

Model lanes

GPT-5.6 Terra (primary): Completed · high confidence · 0 blockers · 1 warning · 0 suggestions
Nemotron 3 Ultra (second opinion): Failed after a partial review · low confidence · 1 blocker · 2 warnings · 0 suggestions

Nemotron output stays in workflow artifacts and does not change the assessment above.

E2E guidance

Advisory only. E2E / PR Gate selects and runs jobs independently.

Recommended E2E: None

1 warning · 0 suggestions

Warnings

Warnings do not block.

`PRA-1` Warning — Pass CVE and NVD data through the correlation contract

Location: .github/workflows/advisory-early-warning.yaml:119
Category: correctness
Problem: The workflow reads `.cveId` from signals and reruns the CLI with `--nvd-records`, but the checked-in signal type has no `cveId` field and the CLI does not parse `--nvd-records`. The CVE list is always empty, so the NVD reconciliation and its rolling-issue annotations never occur.
Impact: The scheduled workflow does not provide the documented supplementary NVD reconciliation for signals that have CVE identifiers. This silently loses the intended informational annotation while reporting a successful workflow.
Recommendation: Add a validated CVE identifier to the signal contract and implement `--nvd-records` consumption in the scan CLI, or remove the unreachable NVD workflow and documentation behavior.
Verification: Inspect `AdvisorySignal` in `scripts/lib/advisory-early-warning.mts` and the accepted flags in `scripts/advisory-early-warning-scan.mts`; neither exposes the fields consumed at workflow lines 119 and 161.
Test coverage: Add a workflow-facing fixture test with a GHSA record containing a CVE ID and an NVD response. Verify that the CLI emits the NVD annotation used by the rolling issue.
Evidence: `.github/workflows/advisory-early-warning.yaml:119` extracts `.cveId` from signal JSON. `.github/workflows/advisory-early-warning.yaml:161-164` passes `--nvd-records` to the scan CLI. `scripts/lib/advisory-early-warning.mts:14-21` defines signals without `cveId`. `scripts/advisory-early-warning-scan.mts:59-76` accepts no `--nvd-records` option.

Workflow run details

This automated review informs maintainers. Warnings and suggestions do not require a response. A maintainer decides whether to merge.

…ce records (#7354) ## Summary Implements the non-operational, code-shaped acceptance criteria of #7338: a correlation library and local CLI that surface public advisories before they reach the reviewed corpus `npm audit` enforces, and provenance records for every audit so disclosure-to-detection timelines are provable from retained artifacts. **Scope note (review follow-up):** per review, the scheduled workflow (`issues: write`) and its operating-policy documentation were split out to #7398, which stays draft until product/security owners record sign-off on #7338's scope / rescan-ownership / alert-destination / response-expectation criterion. This PR defines no operational surface: nothing here is scheduled, holds write permissions, or encodes policy. ## Changes - **Correlation module** (`scripts/lib/advisory-early-warning.mts`, dependency-free): parses GitHub Security Advisory records and correlates them against the reviewed-graph package inventories (`ci/reviewed-npm-audit.json` specs + locked-graph package-locks, alias-aware). Exact npm package + parseable-range matches emit `investigate` signals; ambiguous or CPE-style matches stay `informational` — per the issue, ambiguous matches never block or mutate anything. Unparseable ranges fail safe (noisy, never silent). - **Scan CLI** (`scripts/advisory-early-warning-scan.mts`): local-file-only driver over the module. `--list-packages` emits inventory package names (the input for `/advisories` API queries); `--advisories <file> [--output <file>]` correlates fetched records and prints/writes structured signals. Exits 0 with or without signals; no network or tokens; it does not modify inputs or external state, and `--output` writes the requested local signals file. - **Provenance sidecars**: `scripts/audit-reviewed-npm-graph.mts` (reviewed graphs) and `.github/actions/ci-wechat-runtime-audit/audit.sh` (wechat runtime graph) now write a `*.provenance.json` next to each raw report: scanner identity, npm/node versions, configured registry and derived bulk advisory endpoint, run start/end times, audited package identity, raw report path, extracted GHSA ids — and a `failure` marker whenever the audit attempt itself failed, **including npm's dominant failure mode of parseable error JSON** (verified against a captured dead-registry `npm audit --json` output), so a failed attempt can never masquerade as a clean non-detection. Sidecars ride the existing artifact uploads. - **Docs** (`docs/security/advisory-early-warning.md`): how the correlation works, what each advisory type contributes, CLI usage, and the provenance recorded per audit. Scheduled operation and the proposed policy defaults now live in #7398's version of this page, gated on #7338 sign-off. ## Scope notes - Covered from the issue's acceptance criteria: pre-reviewed-advisory fixture + correlation signal, ambiguous-match non-blocking, reviewed-audit enforcement untouched and authoritative, per-audit provenance incl. actual endpoint (verified empirically: npm ≥7 posts to the bulk endpoint only — the sidecar records no fictitious fallback). - Moved to #7398 (gated on the product/security-owner criterion): scheduled rescans on advisory-data changes, alert routing, and the policy defaults (scope, rescan ownership, alert destination, response SLAs). - Explicitly left as proposals or follow-ups: the npm advisory **database snapshot time** (not exposed by `npm audit`; the sidecar records endpoints + run window instead, and "last comparable non-detection" comes from diffing `advisoryIds` across retained runs), per-upstream-repo advisory polling (needs a package→repo map; named in the doc as the natural extension), Trivy/image-digest provenance, and the #7276 post-mortem classification. ## Testing - 67/67 across the two touched suites (`advisory-early-warning` 49, `reviewed-npm-audit` 18 incl. an integration-style test proving sidecar-write-before-rethrow with a fake npm); tsc strict clean; biome clean; shellcheck/bash -n clean. - Empirical verification on top of unit tests: live GitHub `/advisories` API calls (parameter semantics, pagination truncation repro, `type=malware` `affects` filtering), a captured dead-registry `npm audit` fixture for the failure-marker path, and standalone heredoc runs for the shell sidecar matrix. - Three adversarial local review rounds (premise incl. live-API ground-truthing, correctness incl. empirical repros, contract+security incl. injection tracing); final round dry apart from two Low findings fixed and re-verified before this PR. Refs #7338 Signed-off-by: Dongni-Yang <dongniy@nvidia.com> 🤖 Generated with [Claude Code](https://claude.com/claude-code)  ## Summary by CodeRabbit - **New Features** - Added scheduled “Security / Advisory Early Warning” scan (every 6 hours) with manual trigger support, producing non-blocking correlation signals. - Added structured `npm audit` provenance sidecars (`npm-audit.provenance.json`) next to raw audit output, capturing timing, versions, and extracted advisory identifiers. - **Bug Fixes** - Improved resilience to malformed/non-JSON audit output and strengthened “fails closed” provenance failure reporting. - **Documentation** - Expanded security documentation to describe early-warning correlation logic and provenance artifacts. - **Tests** - Added/updated coverage for early-warning correlation and audit provenance generation behavior.  --------- Signed-off-by: Dongni-Yang <dongniy@nvidia.com> Signed-off-by: Apurv Kumaria <akumaria@nvidia.com> Co-authored-by: Claude Fable 5 <noreply@anthropic.com> Co-authored-by: Apurv Kumaria <akumaria@nvidia.com> Co-authored-by: Carlos Villela <cvillela@nvidia.com>

Resolve to the rebase-equivalent tree after #7354 squash-merged: the branch payload is exactly the scheduled workflow and the doc's scheduled-operation/policy sections, applied on top of main's reviewed #7354 wording (review amendments preserved). Closes #7338 Signed-off-by: Dongni-Yang <dongniy@nvidia.com> Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

github-actions · 2026-07-24T06:20:41Z

🌿 Preview your docs: https://nvidia-preview-pr-7398.docs.buildwithfern.com/nemoclaw

coderabbitai

🧹 Nitpick comments (1)

docs/security/advisory-early-warning.md (1)
6-9: 📐 Maintainability & Code Quality | 🔵 Trivial | ⚡ Quick win

Follow the documentation source conventions.

Put each sentence on one source line and rewrite the changed operational prose in direct, active second-person voice.

docs/security/advisory-early-warning.md#L6-L9: un-wrap the status sentence.

docs/security/advisory-early-warning.md#L77-L87: un-wrap and directly address the reader in the scheduled-operation description.

docs/security/advisory-early-warning.md#L91-L106: un-wrap and directly address the reader in the policy defaults.

As per coding guidelines, “Write documentation in active voice, second person, present tense, and direct language. Keep one sentence per line in Markdown and MDX source files.”
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@docs/security/advisory-early-warning.md` around lines 6 - 9, Update
docs/security/advisory-early-warning.md:6-9, :77-87, and :91-106 so each
sentence occupies one source line and the changed operational prose uses direct,
active second-person, present-tense language; apply the requested wording
changes to the status sentence, scheduled-operation description, and policy
defaults.
Source: Coding guidelines

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Nitpick comments:
In `@docs/security/advisory-early-warning.md`:
- Around line 6-9: Update docs/security/advisory-early-warning.md:6-9, :77-87,
and :91-106 so each sentence occupies one source line and the changed
operational prose uses direct, active second-person, present-tense language;
apply the requested wording changes to the status sentence, scheduled-operation
description, and policy defaults.

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Enterprise

Run ID: 784aaca6-ac9e-4d01-875c-1a321ec0e424

📥 Commits

Reviewing files that changed from the base of the PR and between ccf3dd4 and 9cf4f42.

📒 Files selected for processing (2)

.github/workflows/advisory-early-warning.yaml
docs/security/advisory-early-warning.md

Resolve to the restructured non-operational slice: the NVD reconciliation library, scan-CLI plumbing (--nvd-records, --inventory, cveId), tests, and the reconciliation-semantics + #7276 post-mortem docs, rebased onto main after #7354 squash-merged. The scheduled workflow's NVD step and the policy-defaults doc section move to the sign-off-gated #7398, matching the #7354/#7398 split. Refs #7338 Signed-off-by: Dongni-Yang <dongniy@nvidia.com> Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

The operational half of the NVD reconciliation split (#7370 carries the library and CLI plumbing): the scheduled workflow queries NVD for up to 20 signal CVE ids per run (7 s spacing for the unauthenticated rate limit, cap logged when exceeded) and appends the agreement annotation to the rolling-issue lines. NVD being unreachable degrades to an "NVD: unavailable" note and never fails the run — deliberately softer than the fail-loud GitHub advisory fetch, since NVD is supplementary. Annotations extend line text only; the dedupe key is unchanged. Merge order: #7370 must land first — this workflow invokes the scan CLI's --nvd-records flag, which #7370 introduces. Refs #7338 Signed-off-by: Dongni-Yang <dongniy@nvidia.com> Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

coderabbitai

Actionable comments posted: 1

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Inline comments:
In @.github/workflows/advisory-early-warning.yaml:
- Around line 101-153: Wire NVD metadata through
scripts/lib/advisory-early-warning.mts and
scripts/advisory-early-warning-scan.mts. Add each signal’s cveId (and required
NVD metadata) to the emitted signal shape, accept and parse the workflow’s
--nvd-records argument, and merge matching NVD records into the signals so the
existing workflow worklist and issue-body NVD notes are populated.

🪄 Autofix (Beta)

Fix all unresolved CodeRabbit comments on this PR:

Push a commit to this branch (recommended)
Create a new PR with the fixes

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Enterprise

Run ID: 7e49dab4-7fe8-41da-b2e9-2720f2b5826d

📥 Commits

Reviewing files that changed from the base of the PR and between 9cf4f42 and 045b81c.

📒 Files selected for processing (2)

.github/workflows/advisory-early-warning.yaml
docs/security/advisory-early-warning.md

🚧 Files skipped from review as they are similar to previous changes (1)

docs/security/advisory-early-warning.md

apurvvkumaria

Blocking pending design alignment.

This PR should not merge until product/security owners complete the design evaluation for the scanning tool, including alignment with the parallel OpenShell work. The scheduled workflow, supported-image scope, ownership model, alert destination, and response expectations need one agreed operating design before NemoClaw commits to this implementation.

Please keep this PR unmerged until that decision is recorded on #7338, then update the implementation and documentation to match the agreed design. This is an operational-design blocker, not a rejection of the underlying early-warning goal.

…-mortem (#7370) ## Summary The **non-operational NVD slice** of #7338: NVD as a supplementary reconciliation source (a pure library plus offline CLI plumbing) and the **#7276 post-mortem** answering the issue's "Questions to answer" — the post-mortem is an explicit #7338 acceptance checkbox. **Restructured 2026-07-24** to match the #7354/#7398 split: this PR was previously stacked on pre-narrowing #7354 and included the scheduled workflow's NVD step. That step (and the policy-defaults doc section) now belongs to the sign-off-gated #7398; this PR is rebased onto main and carries only pieces that define no operational surface — nothing here is scheduled, holds write permissions, or encodes policy. It can merge on normal review. ## Changes - **`scripts/lib/nvd-reconciliation.mts`** (new, pure, dependency-free): parses NVD 2.0 API records (malformed → null, never throws; live-verified against `services.nvd.nist.gov`, including a 358-criteria CPE record and the reserved-id empty envelope) and reconciles each signal: `corroborated` / `nvd-missing` / `nvd-divergent`. Annotations are **informational only** — they carry no action or confidence and cannot alter a signal's dedupe identity (test-asserted), per the issue's requirement that ambiguous NVD/CPE matches never become authoritative npm mappings. - **Signal plumbing**: `AdvisorySignal` gains an optional `cveId` from the GHSA record's `cve_id`; the scan CLI accepts `--nvd-records <file>` (offline attachment — the CLI never performs network requests) and `--inventory <file>` (hermetic test runs). - **Docs**: NVD reconciliation semantics, and the #7276 post-mortem classifying each finding per the acceptance criteria — `fast-uri`: reviewed-mapping delay, directly demonstrated; `@opentelemetry/core`: audit-coverage/execution-order gap; Jaeger: consistent but unproven; `tar`: unproven, evidence not retained. Post-mortem wording states plainly which covering mechanisms are merged and which ride the gated #7398. ## What moved to #7398 The scheduled workflow's NVD reconciliation step (rate-capped fetch, `NVD: unavailable` degradation, rolling-issue annotation) and the proposed policy defaults — operational surface under #7338's owner-sign-off gate. ## Documentation Writer Review - [x] Documentation writer subagent reviewed the completed changes - Result: `no-docs-needed` - Evidence: `scripts/lib/nvd-reconciliation.mts` and `test/nvd-reconciliation.test.ts` now traverse nested NVD configuration nodes as an internal correctness fix for the documented vulnerable-CPE parsing contract; no user-facing command, flag, configuration, output schema, or workflow changes. - Agent: Codex Desktop   ## Testing - `test/nvd-reconciliation.test.ts`: 122 tests green across the three suite files locally; tsc strict clean on both `.mts` entry points; biome, SPDX, test-size-budget, title-style, and source-shape guardrails verified locally. - CLI smoke-verified: `--list-packages` runs against the repo inventory. - Maintainer follow-up: `npx vitest run --project integration test/nvd-reconciliation.test.ts test/advisory-early-warning.test.ts` passes 93/93 with nested NVD configuration coverage. Refs #7338 (the issue closes via #7398 once product/security owners record sign-off) Signed-off-by: Dongni-Yang <dongniy@nvidia.com> 🤖 Generated with [Claude Code](https://claude.com/claude-code)  ## Summary by CodeRabbit * **New Features** * Advisory early-warning signals can include an optional CVE identifier when available. * Added offline NVD supplementary reconciliation that annotates results with corroborated, nvd-missing, or nvd-divergent. * Added `--inventory <file>` for hermetic inventories and `--nvd-records <file>` to attach pre-fetched NVD reconciliation data without network access. * **Documentation** * Expanded early-warning correlation documentation, including NVD reconciliation behavior and the updated `#7276` post-mortem. * **Bug Fixes** * Malformed inventory entries now fail the run instead of silently shrinking the inventory. * **Tests** * Updated and expanded coverage for CVE extraction, NVD processing, and CLI offline behavior.  --------- Signed-off-by: Dongni-Yang <dongniy@nvidia.com> Signed-off-by: Apurv Kumaria <akumaria@nvidia.com> Co-authored-by: Claude Fable 5 <noreply@anthropic.com> Co-authored-by: Apurv Kumaria <akumaria@nvidia.com>

apurvvkumaria · 2026-07-24T20:52:02Z

Closing per maintainer direction: we are not adding scanning to the NemoClaw repository. This PR will not be merged.

Dongni-Yang and others added 6 commits July 22, 2026 11:32

docs(security): add SPDX header to advisory early-warning doc

1b53bb7

CodeRabbit: every source file carries the SPDX header; match the sibling docs/security file comment syntax. Refs #7338 Signed-off-by: Dongni-Yang <dongniy@nvidia.com> Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

Merge remote-tracking branch 'origin/main' into dongniy/7338-audit-pr…

9759ce2

…ovenance

Merge remote-tracking branch 'upstream/main' into dongniy/7338-audit-…

c52ee04

…provenance

wscurran added area: ci CI workflows, checks, release automation, or GitHub Actions area: security Security controls, permissions, secrets, or hardening feature PR adds or expands user-visible functionality needs: unblock Blocked item needs dependency or decision resolved labels Jul 23, 2026

Base automatically changed from dongniy/7338-audit-provenance to main July 23, 2026 22:39

Dongni-Yang marked this pull request as ready for review July 24, 2026 06:20

Dongni-Yang assigned Dongni-Yang and unassigned Dongni-Yang Jul 24, 2026

Dongni-Yang added the v0.0.94 label Jul 24, 2026

coderabbitai Bot reviewed Jul 24, 2026

View reviewed changes

prekshivyas added v0.0.95 and removed v0.0.94 labels Jul 24, 2026

coderabbitai Bot reviewed Jul 24, 2026

View reviewed changes

Comment thread .github/workflows/advisory-early-warning.yaml

senthilr-nv assigned apurvvkumaria Jul 24, 2026

apurvvkumaria requested a review from senthilr-nv July 24, 2026 16:46

apurvvkumaria requested changes Jul 24, 2026

View reviewed changes

apurvvkumaria removed the v0.0.95 label Jul 24, 2026

coderabbitai Bot mentioned this pull request Jul 24, 2026

docs: complete post-tag audit fixes #7495

Merged

26 tasks

apurvvkumaria closed this Jul 24, 2026

Navigation Menu

Uh oh!

feat(security): schedule the advisory early-warning scan - #7398

feat(security): schedule the advisory early-warning scan#7398
Dongni-Yang wants to merge 8 commits into
mainfrom
dongniy/7338-early-warning-workflow

Dongni-Yang commented Jul 22, 2026 •

edited by coderabbitai Bot

Loading

Uh oh!

copy-pr-bot Bot commented Jul 22, 2026

Uh oh!

coderabbitai Bot commented Jul 22, 2026 •

edited

Loading

Walkthrough

Changes

Sequence Diagram(s)

❌ Failed checks (1 inconclusive)

Uh oh!

github-code-quality Bot commented Jul 22, 2026 •

edited

Loading

Uh oh!

github-actions Bot commented Jul 22, 2026 •

edited

Loading

Warnings

`PRA-1` Warning — Pass CVE and NVD data through the correlation contract

Uh oh!

github-actions Bot commented Jul 24, 2026

Uh oh!

coderabbitai Bot left a comment

Uh oh!

coderabbitai Bot left a comment

Uh oh!

Uh oh!

apurvvkumaria left a comment

Uh oh!

apurvvkumaria commented Jul 24, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

Uh oh!

Conversation

Dongni-Yang commented Jul 22, 2026 • edited by coderabbitai Bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Why this is separate

Changes

Merge checklist

Summary by CodeRabbit

Uh oh!

copy-pr-bot Bot commented Jul 22, 2026

Uh oh!

coderabbitai Bot commented Jul 22, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Walkthrough

Changes

Sequence Diagram(s)

❌ Failed checks (1 inconclusive)

Uh oh!

github-code-quality Bot commented Jul 22, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Code Coverage Overview

TypeScript / code-coverage/plugin

TypeScript / code-coverage/cli

Uh oh!

github-actions Bot commented Jul 22, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

PR Review Advisor — Informational

Model lanes

E2E guidance

Warnings

PRA-1 Warning — Pass CVE and NVD data through the correlation contract

Uh oh!

github-actions Bot commented Jul 24, 2026

Uh oh!

coderabbitai Bot left a comment

Choose a reason for hiding this comment

Uh oh!

coderabbitai Bot left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

apurvvkumaria left a comment

Choose a reason for hiding this comment

Uh oh!

apurvvkumaria commented Jul 24, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

Dongni-Yang commented Jul 22, 2026 •

edited by coderabbitai Bot

Loading

coderabbitai Bot commented Jul 22, 2026 •

edited

Loading

github-code-quality Bot commented Jul 22, 2026 •

edited

Loading

github-actions Bot commented Jul 22, 2026 •

edited

Loading

`PRA-1` Warning — Pass CVE and NVD data through the correlation contract