Hi all

Present

Key Outcomes

• PSA Crypto/mbed TLS maintenance is at risk; current maintainers plan to stop development/maintenance by April 27, and the project is actively seeking new contributors or organizations to take over. 
• PSA Crypto API support is progressing: mbed TLS 3.6 LTS supports PSA Crypto 1.1, and mbed TLS 4.1 supports 1.2 (with some gaps); full 1.5 support is in progress. 
• PQC support is moving forward: ML-DSA is implemented via a fork of the ML-DSA native project from the PQC Alliance, with PSA Crypto API support expected by the end of the year. 
• TF security incident handling is being streamlined to reduce overhead and speed up fixes, including a proposal to decouple CVE creation from vulnerability fixing and move toward a more kernel-like model. 
• Zephyr is proposing to end embargoes once a fix is merged into Git, as AI tools make embargoes less effective. 

Decisions Made

• PSA Crypto 1.2 will not be backported to mbed TLS 1.x; a separate version was released for that branch.
• Security fixes in TF PSA Crypto 1.2 will be backported to the 1.x LTS branch. 
• TF PSA Crypto ML-DSA implementation will leverage the ML-DSA native project from the PQC Alliance rather than a from-scratch implementation. 
• TF PSA Crypto will only support ML-DSA-87 initially
• TF PSA Crypto driver repository has moved to the Mbed TLS GitHub organization and is open for vendors to push their PSA Crypto drivers. 

Progress & Status

• mbed TLS maintenance transition: Board members and downstream consumers are checking for funding or personnel within their organizations; some positive queries have been received, but no solid agreements yet. 
• ML-DSA: The team is actively working on ML-DSA support via the PSA Crypto API this quarter. 
• Security issues: The bug bounty program has been paused, and the volume of new incoming issues is decreasing, though a considerable backlog remains. 
• mbed TLS 3.6 LTS will reach end-of-life in March 2027; users are encouraged to migrate to 4.x and PSA Crypto APIs. 
• Chinese crypto algorithms (SM2, SM3, SM4) are being investigated for potential upstreaming via PSA API or driver interfaces. 
• Linaro is expected to become a CNA, which may provide more reliable CVE creation for TF projects. 

Risks & Blockers

• Maintenance risk: If no new maintainers step up by April 27, mbed TLS and X.509 libraries may stop receiving security fixes and regular updates. 
• PR review capacity: PR review remains a challenge for the PSA Crypto project; more reviewers are needed. 
• Security process overhead: High volume of reports, complex triage, and manual CVE/advisory management are slowing down fixes. 
• CRA implications: The Cyber Resilience Act (CRA) may impose reporting requirements on open-source stewards, including a 24-hour deadline for actively exploited vulnerabilities, which is seen as unrealistic for volunteer-led projects. 
• AI-generated reports: AI tools are increasing the volume of reports, some of which are low-quality or duplicates, adding to the triage burden. 

Pending Confirmation

• New maintainers for mbed TLS and X.509 libraries need to be identified and committed before the April 27 deadline. 
• ML-KEM is on the roadmap but work has not yet started; it will likely follow the completion of ML-DSA. 
• CRA requirements for open-source stewards remain unclear and are expected to be resolved by December of next year. 
• Proposed security process changes need further review from maintainers.

Action Items

• Shebu: Collate responses from organizations regarding mbed TLS maintenance and have Bharath present them at the next board meeting. 
• Dan: Create a PR for the proposed security incident handling process changes and send it for formal review by the TSC and project security teams. 
• Mbed TLS users: Consider moving to mbed TLS 4.x and PSA Crypto 1.1 before the 3.6 LTS end-of-life in March 2027. 
• HW Vendors: Encouraged to push their PSA Crypto drivers to the new TF PSA Crypto driver repository.