Org (tenant) layer
- New Org model with org_uuid denormalized onto workspaces, assets,
vulnerabilities and runs so cross-workspace queries need no join
- Automatic attribution via BeforeAppendModel hooks; importers stay org-unaware
- Read semantics: empty org means no filter (backward compatible)
Write semantics: empty org coerced to the default org
- Migration backfills every pre-existing row into the default org
- CLI: osmedeus org create/show/assign/use/rename/delete
- API: /osm/api/orgs CRUD plus ?org= on assets, vulns, runs and workspaces
npm distribution
- npm install -g @j3ssie/osmedeus ships the Go binary through npm
- One npm name with version-suffixed platform builds pulled in as aliased
optionalDependencies, so an install downloads exactly one binary
- Binary ships gzipped and is decompressed on first run into a
version-scoped cache, so an upgrade can never exec a stale binary
- make bump-version is the single source of truth for the version constant
Bundled agent skills
- public/skills/ embedded in the binary, installed via osmedeus skills install
- Filesystem-driven discovery: a new bundle needs no code change
- make sync-skills mirrors bundles out to the standalone skills repo
Platform sub-projects
- Vendor dashboard, registry and workflow under platform/ so they version
with the engine they talk to; make sync-platform publishes them out
- Rebuild the embedded UI in public/ui/
- Track tmux sessions in ActiveRun and kill them on run cancel via new
TmuxHooks indirection; expose killed_tmux_sessions in CLI and API responses
- Add ExecuteSSHCommand with remote pidfile + process-group kill watcher so
cancelling a run actually terminates remote scans (not just the local session)
- Route ssh_exec/ssh_rsync/sync_* through the run's cancellable context via
new RunContextHooks
- Switch docker-publish to sequential per-arch buildx builds + imagetools
manifest to avoid OOM on multi-arch builds; add docker-buildx-setup target
- Cross-compile Dockerfile via BUILDPLATFORM/TARGETOS/TARGETARCH and retry
SAST binary installs to survive QEMU-flaky downloads
- Bump version to v5.0.3
- Implement cloud provider infrastructure (DigitalOcean, AWS, GCP, Linode, Azure) with Pulumi integration for distributed scanning
- Add nmap and tmux utility functions for port scanning results processing and long-running background session management
- Introduce webhook-triggered run execution with unique UUID and authentication key support for external integrations
- Add Docker Compose distributed-e2e.yaml with Redis, master, and worker services
- Add Make targets: distributed-e2e-up, distributed-e2e-run, distributed-e2e-down
- Replace zap logger with terminal printer for consistent CLI output in master/worker
- Add Redis connection info and task details to startup and lifecycle logs
- Fix DNS records field: use 'dns_records' key with backward compat for 'a' key
- Add optional asset_type and source parameters to db_import_custom_asset
- Update Swagger docs for LLM embeddings schema
- Set run_uuid tracking in worker task execution for distributed runs
- Register distributed database hooks and ensure external binaries in PATH
- Improve terminal formatting with colors for worker/master IDs and log messages
- Add WorkflowHelp struct with Usage and ExampleTargets for CLI documentation
- Add Optional field to Artifact model and database schema with migration support
- Implement workflow search functionality by name, description, and tags in CLI list command
- Add --usage and --search flags to workflow list command with multiple filtering options
- Display workflow usage info in show command when Help is defined
- Support help inheritance in workflow extends/inheritance resolver
- Update vulnerability counters from database after SARIF imports
- Add comprehensive Help unit tests covering parsing, cloning, and mutation isolation
- Improve test helpers with streaming output, diagnostics, and file validation utilities
- Add fourth general canary test for domain-list-recon flow with artifact validation
- Extract 'owner__repo' slugs from code hosting platforms (GitHub, GitLab, Bitbucket, Codeberg) for workspace naming preference
- Implement magic byte detection for archive formats in extractTo() function to support headerless archives
- Implement GetRunSteps() and GetRunArtifacts() API endpoints with database queries
- Format code for consistency in constants, type mappings, and test files
- Convert main Dockerfile to multi-stage build: compile from source in builder stage, runtime on minimal debian-slim image
- Replace remote install.sh with local source binary compilation using Go 1.25
- Pre-install SAST tools (trivy, semgrep, kingfisher, bearer) in both Dockerfile and canary variant
- Fix semgrep installation on Debian by adding --break-system-packages flag to pip install
- Simplify canary Dockerfile by copying pre-built binary and adjusting install order for clarity
- Replace curl-based installation with 'osmedeus install base --preset' command
- Add explicit binary installation for trivy, semgrep, and kingfisher for SAST testing
- Comment out original install script line for reference
- Add docker-compose.canary.yaml with multi-stage Dockerfile to build osmedeus from source and layer onto toolbox base, ensuring canary tests exercise current code not released binaries
- Implement three canary test scenarios (repo SAST scan, domain-lite reconnaissance, CIDR IP scanning) with lifecycle management (container startup/cleanup) and comprehensive assertions on filesystem artifacts and API records
- Add Makefile targets (test-canary-all, test-canary-repo, test-canary-domain, test-canary-ip, canary-up, canary-down) for granular test execution with configurable timeouts (20-60 minutes)
- Add extract_to() utility function for auto-detecting archive formats (.zip, .tar.gz, .tar.bz2, .tar.xz, .tgz) and extracting with automatic destination cleanup
- Refactor Dockerfile to use slim debian base with install script, reducing image complexity and improving build reproducibility
- Update Dockerfile.dev to match production environment (Go 1.25 with full toolchain) while adding preset workflow initialization
- Update trivy binary registry metadata from v0.69.0 to v0.69.1 with direct platform-specific download URLs replacing dual curl installation script
- Remove unnecessary goreleaser skip flags from manual-release workflow
- Add local-release make target for testing macOS and Linux ARM64 builds
- Refactor Dockerfile.toolbox to use direct-fetch with Ubuntu 24.04 base
- Create Dockerfile.toolbox-nix for Nix-based installation variant
j3ssie