ASD ISM 2026 Scan software artefacts for malicious content
What it says
All software artefacts are scanned for malicious content before being imported into the authoritative source for software.
Quoted as published. Everything else on this page is written by Control Stack.
In plain English
Every software artefact, from in-house files to third-party libraries, must be scanned for malicious content before it enters the authoritative source for software.
All software artefacts must be scanned for malicious content before they are imported into the authoritative source for software, the trusted store the organisation keeps for its code and the files used to build it (ISM-2023). Artefacts include third-party libraries, compiled code, configuration files and any other file development uses, whether made in-house or elsewhere. Malicious content is anything deliberately harmful, such as a backdoor or an install script that downloads malware. Testing for known weaknesses is a separate control, ISM-2028.
Free. Save your status, tasks and notes.
Why it matters
A library from a public registry, a supplier delivery or another team's file can carry malicious code planted for later use. Once inside the authoritative source, it reaches every build that uses it, so scanning before import gives a chance to catch it first.
How to implement
List every route by which artefacts reach the authoritative source.
Include package, container image and AI model pulls, supplier deliveries, uploads of build outputs and configuration files, every push to source repositories, and files brought in from other teams. Check the repository's own audit logs for write activity, so routes nobody documented also show up.
Put a malicious content scan in front of each route.
Use the repository manager's quarantine so pulled packages and images wait for scanner clearance, scan uploads in staging before promoting them, and scan pushes in a server-side pre-receive hook. Choose scanners that detect known malware and signs of malicious packages, such as obfuscated code or install scripts that fetch payloads.
Make each scan fail closed.
Configure the gate so an artefact whose scan errors, times out or is skipped is refused rather than imported. Hold anything flagged in quarantine, and import it only if an investigation shows the detection was false.
Remove routes that bypass the scan.
Take away direct write access that skips the gate, such as manual administrator uploads, or send it through the same scan. Where a hosted repository offers no scanning hook, import through a scanned staging repository you control.
Keep it running
Keep scanner signatures and detection rules updated, and follow up scans that error or time out. Recheck the gates whenever a new package ecosystem, registry, artefact type or import route is added.
Evidence and audit
- Check 1 of 3What an auditor asks for
Ask for
The repository's audit logs of writes for a recent period, with the documented import routes and the scan on each.
Look at
Whether every route that actually wrote artefacts into the authoritative source passed them through a malicious content scan first.
Good evidence
Every route seen in the logs, including administrator uploads, is documented and scans artefacts before import, and no artefact arrived by an unscanned route.
- Check 2 of 3What an auditor asks for
Ask for
A demonstration sending a harmless test file the scanner is set to detect through each import route, and an import attempted while the scanner is unavailable, in a test instance or under change control.
Look at
Where each attempt is stopped.
Good evidence
The test file is refused or quarantined on every route before it reaches the authoritative source, and nothing is imported while the scanner is unavailable.
- Check 3 of 3What an auditor asks for
Ask for
Scan records for a sample of artefacts picked from the authoritative source itself, covering source code, third-party libraries, compiled code and configuration files.
Look at
Whether each sampled artefact has a malicious content scan dated before its import.
Good evidence
Every sampled artefact, of every type, has a scan recorded before its import date, and nothing flagged was imported without an investigation showing a false detection.
Mappings
Closest match in ISO/IEC 27001: A.5.21, Managing information security in the information and communication technology (ICT) supply chain, which this control partially meets.
ISO/IEC 27001: 3 mappings
- A.5.21Managing information security in the information and communication technology (ICT) supply chainPartially meets
This control partially meets A.5.21.
Annex A 5.21 covers defining and implementing processes and procedures for the information security risks of the ICT products and services supply chain. ISM-2026 is one specific case within that scope, an operational check at the point of import, requiring every software artefact to be scanned for malicious content before entering the authoritative source for software.
This control partially meets A.8.7.
ISM-2026 requires all software artefacts to be scanned for malicious content before they are imported into the authoritative software source. This relates directly to ISO/IEC 27001:2022 Annex A 8.7, which requires organisations to implement protection against malware, including measures such as scanning to prevent malicious code introduction. Annex A 8.7 is broader, covering organisation-wide anti-malware controls (e.g. endpoints, email, web, user awareness), while ISM-2026 is a specific supply-chain/gatekeeping control on software artefact ingestion.
This control partially overlaps A.8.29.
ISM-2026 requires scanning software artefacts for malicious code before they are imported into the authoritative software source. Annex A 8.29 relates by requiring security testing to be defined and implemented within the development and acceptance lifecycle, which can include malware and integrity checks on build outputs and dependencies. Annex A 8.29 is broader (covers multiple kinds of security testing and SDLC stages), while ISM-2026 is specifically focused on malicious code scanning at the import/gating point.
Mappings show how controls relate across frameworks. They do not imply full equivalence or certification.




