OpenShell
OpenShell is the safe, private runtime for autonomous AI agents.
Is OpenShell safe? Promising trust profile, but some evidence still deserves review.
Compare OpenShell
How does it stack up against its Sandboxes & Runtimes neighbours?
Pick any agent to compare →In detail: OpenShell scores 74.1/100 (Grade B), ranked #271 of 1371 tracked open-source AI agent projects, on evidence coverage B (3 of 5 independent signal types). The public evidence: no package-provenance attestation found; OSSF Scorecard rates its supply-chain practices 6.5/10; 100% of recent commits are signed; no published advisory affects its latest release (openshell); last pushed 2026-10-14. Every point is earned from checkable signals — never paid placement. How scoring works →
What OpenShell can reach, and who runs it
Checked 2026-10-09, each with its source. PUBLISHER is the company's or project's own statement, which we haven't verified; CHECKED is something we observed ourselves. These facts describe the product. They don't change its score.
Who runs it
- CHECKED Made by NVIDIA; open source under Apache-2.0.
What it can reach
- PUBLISHER Each agent runs in an isolated sandbox: kernel controls confine which files and system calls it can use, and every network connection passes through a policy.
Where your data goes
- PUBLISHER Sends anonymous telemetry limited to operational categories and counts, not sandbox names, hostnames, file paths, prompts, credentials or user content. OPENSHELL_TELEMETRY_ENABLED=false turns it off.
Security practice
- CHECKED The repository has a SECURITY.md and GitHub private vulnerability reporting is enabled.
How OpenShell could raise its score
Each line changes one public signal and recomputes with the live scoring function. The gains don't add up exactly, because the score is capped near the top.
- Publish build provenance for its packagesPyPI Trusted Publishing attestations +14.7 → 88.8 A · #1 of 11
- Raise the OSSF Scorecard from 6.5 to 9.0lowest checks: CII-Best-Practices 0, Fuzzing 0, Signed-Releases 0 +5.2 → 79.3 B · #2 of 11
Chain of custody
Scanners check what the code says. This traces who ships OpenShell and whether that has changed: from the source repository, through how changes are reviewed and released, to the package you install. These are the checks for OWASP ASI04 Agentic Supply Chain Vulnerabilities.
-
Source Signed
github.com/NVIDIA/OpenShell, Apache-2.0, last pushed 2026-10-14. 100% of the last 100 commits carry a verified signature, so changes trace to a verified account.
-
Review and release Mixed
OpenSSF Scorecard rates the repository's practices 6.5/10 (scanned Sep 29, 2026, refresh pending). The checks that decide who can get a change released:
- Code Review9
- Branch Protection4
- Signed Releases0
- Dangerous Workflow10
- Token Permissions0
- Pinned Dependencies6
All 18 Scorecard checks
Binary-Artifacts 10Branch-Protection 4CI-Tests 10CII-Best-Practices 0Code-Review 9Contributors 10Dangerous-Workflow 10Dependency-Update-Tool 10Fuzzing 0License 10Maintained 10Packaging 10Pinned-Dependencies 6SAST 10Security-Policy 10Signed-Releases 0Token-Permissions 0Vulnerabilities 0 -
Published packages Not linked
- PyPI openshell Source link isn't a GitHub repo No build attestation
Without a build attestation, nothing proves the published files were built from this repository.
-
What you install today No known advisories
Latest release checked: openshell 0.1.3. No published advisory affects it (OSV, checked 2026-10-13).
Changes to this chain
No change to package provenance, package source links, Scorecard coverage or license in HVTracker's daily snapshots of OpenShell.
Evidence behind this score
Coverage B: 3 of 5 independent evidence types found.
- GitHub repository data
- Package downloads
- Supply-chain checks
- Public actions (missing)
- Community mentions (missing)
Verify this score yourself
Every build re-issues OpenShell’s score as an Ed25519-signed credential, valid for 7 days. You can check it offline against HVTracker’s published key; if anyone changes a number after signing, verification fails.
What this score doesn’t check
It covers who ships the code, not what the code does. It doesn’t read tool descriptions or prompts for injected instructions, watch runtime behaviour, or find bugs nobody has disclosed yet. For that, run a content scanner before you connect it, such as Cisco MCP Scanner or Snyk Agent Scan.
How OpenShell compares in Sandboxes & Runtimes
- #1 BoxLite 85.0 +10.9
- #2 E2B 77.7 +3.6
- #3 OpenShell 74.1 this agent
- #4 Cube Sandbox 71.7 −2.4
- #5 OpenSandbox 70.7 −3.4
Bars show each HVTrust score; the tick marks OpenShell’s 74.1.
Where the 74.1 comes from
HVTrust dimensions vs the Sandboxes & Runtimes average
74.1 / 100 · 100.0% confidenceOpenShell Sandboxes & Runtimes average (11 agents)
Quick Trust Read
How to read this: HVTrust (0–100) weighs supply-chain signals (provenance, OSSF Scorecard, signed commits, open license) alongside real-world adoption. Grade B reflects the trust score band: A ≥ 80, B ≥ 65, C ≥ 50, D < 50. Evidence coverage B is separate — it grades how many independent signal types back the score (3 of 5), so a high score on thin evidence stays visible. Full methodology →
Rank Trend
Activity & Reach
Analysis
Activity Inputs
90.3 / 100Common questions about OpenShell
Does OpenShell publish package provenance?
Does OpenShell have an OpenSSF Scorecard?
Is OpenShell actively maintained?
What license does OpenShell use?
Are OpenShell's commits signed?
Not a safety endorsement. HVTracker describes what public signals show, not whether a project is safe for your use case. Run your own security review before adopting in production.
Compare OpenShell head-to-head
AI agent surface
MCP, providers, tool surface
These runtime-trust fields — detected from public repo docs and manifests — contribute a bounded adjustment to this project's HVTrust score alongside supply-chain evidence. The exact values each field can add or subtract are documented in the methodology → Compare this surface across every listed agent in the capability matrix →
- MCP signal live
- External deps live
- Tool / plugin surface live
- Package provenance drift live
Detected changes to OpenShell's runtime surface and supply-chain posture, from daily public-signal snapshots. A change here means our detectors see something different — a genuinely changed capability, or better evidence of an existing one.
Maintain OpenShell?
For maintainers
HVTrust scores OpenShell from public signals only — we never contact maintainers first. If a signal is wrong, stale, or missing (provenance you publish, a Scorecard you run, signed releases), tell us and we'll review it. Corrections are public and tracked on GitHub.
Reputation Timeline
Signal history
Embed Badge Badge guide for maintainers →
For maintainers
[](https://hvtracker.net/agents/openshell)
<a href="https://hvtracker.net/agents/openshell"><img src="https://hvtracker.net/badge/openshell.svg" alt="HVTrust"></a>
Other agents in Sandboxes & Runtimes
GitHub REST API (repo, commits, stars, forks, license) · PyPI / pypistats (downloads, provenance) · OpenSSF Scorecard CLI
Each agent's signals refresh once daily across 6 staggered batches. Methodology v4.4 · Raw JSON