Security fixes are provided for the latest 3.0 release line. Older releases are not maintained — please update before reporting.

Supported release line: 3.0.x

Report security issues through the project's GitHub issue tracker:

https://github.com/aagedal/Aagedal-Photo-Agent/issues/new

Start the issue title with [security] so it can be triaged and labelled quickly. Aagedal Photo Agent is a local, single-user desktop app — it runs no servers, so coordinated public disclosure here is appropriate for most findings.

To avoid handing out a working exploit before a fix is available, please:

• In the initial report, describe the issue and its impact at a level that lets it be assessed, but hold back a full working exploit or a ready-to-run malicious sample if the flaw looks exploitable and unpatched. A maintainer will reply in the thread to arrange the remaining details.
• Include the app version (Settings → Licenses shows the build) and your macOS version.
• Attach a minimal reproduction (a sample file or configuration) once asked, or from the start if the issue is low-risk.

GitHub issues in this repository are public. If you believe a finding is too sensitive to describe in public at all, open a minimal issue titled [security] request private contact with no technical detail, and a maintainer will arrange a channel.

You can expect an acknowledgement within 7 days and a status update within 30 days. Once a fix ships, credit is given in the changelog unless you prefer to remain anonymous. As this is a small open-source project there is no bug-bounty program.

Aagedal Photo Agent is a local, single-user macOS application. The most relevant areas to look at:

• Untrusted file parsing. The app reads images, video/audio metadata, EXIF/IPTC/XMP metadata, PDFs, and .xmp / JSON / face-data sidecars that may originate from untrusted sources. Parsing uses the pure-Swift SwiftMediaMetadata package, Apple ImageIO and media frameworks, the app's own sidecar decoders, and bundled command-line helpers where applicable. Memory-safety, command-injection, path-handling, or parser-confusion issues triggered by a malicious file are in scope.
• Path traversal. Imports that read paths from files — e.g. keyword-list and Known People / Teams database (ZIP) import — must stay within their intended directories.
• Credential and key storage. FTP/SFTP passwords and C2PA signing private keys are stored in the macOS Keychain and must never be written to logs, sidecars, exported metadata, or synced via iCloud.
• Content credentials (C2PA). Signing certificate/key handling and manifest generation.
• Network paths. FTP/FTPS/SFTP upload, geocoding and map lookups, and the Sparkle auto-update channel (appcast over HTTPS, releases verified with an EdDSA signature).
• iCloud sync. Synced categories must never include passwords, signing keys, certificates, or machine-specific paths.

• Attacks requiring a compromised local account or physical access to an unlocked Mac.
• Denial of service from intentionally malformed files that does not lead to code execution or data disclosure (please still report crashes — they are fixed as ordinary bugs).
• Findings in third-party dependencies that are already public and fixed upstream; please report those upstream and reference them here.

This software is distributed under the GPL-3.0 with no warranty (see LICENSE, sections 15–16). The security process above is a good-faith effort, not a contractual guarantee.