Skip to content

Command Palette

Search for a command to run...

Version 1.2.0

Security scanning

Local, read-only dependency vulnerability scanning for Go and Rust apps against OSV.dev advisories, with evidence-classed findings, scheduled scans, and an exit-code CI gate.

phelix security inventories the Go and Rust dependencies of an application and evaluates them against OSV.dev advisories, fully locally: dependency files are parsed in place, nothing is built, and nothing is downloaded. Go dependency versions are resolved authoritatively with go list -m (proxy disabled: resolution uses the local module cache only, so a warm cache from normal builds answers offline); without a usable toolchain the scan degrades to manifest evidence and says so.

Evidence classes

Findings carry their evidence class, and the class is part of the report rather than a footnote:

ClassMeaning
affectedThe resolved version is established (Cargo.lock, a pinned replace directive, or the resolved Go module graph) and lies inside the affected range.
potentially_affectedOnly manifest minimums are known; the reasoning is attached so you can decide.
unknownThe scan could not decide. It is reported rather than silently cleared, and reachability is always unknown.

The exit 25 gate

A scan exits 25 when findings reach the configured threshold, which makes the command a usable CI gate. An offline scan that could not check every package reports advisory_coverage: partial and never presents itself as clean. When OSV.dev is unreachable entirely, the scan exits 30 instead, so “findings found” and “could not check” never share an exit code.

Commands

Terminal
phelix security scan
phelix security scan myapp
phelix security report myapp --previous
phelix security history myapp --limit 20
phelix security show GHSA-xxxx-xxxx-xxxx
phelix security config myapp
CommandWhat it does
phelix security scan [app]Scan and persist a record; progress renders on stderr and --json stays pure.
phelix security report [app]Latest completed scan; --previous diffs it against the scan before it.
phelix security history [app]Recorded scans with status and counts.
phelix security show <ID>One advisory, cache first.
phelix security config [app]Effective settings; --set-enabled, --set-scan-interval, and --set-fail-threshold edit phelix.yaml safely.

Run without arguments in a terminal, the commands guide you: phelix security opens a menu, and scan, report, history, show, and config ask for a target application. Scripts and CI keep the exact non-interactive behavior, and every command accepts --json.

Configuration

Scheduled scans run opt-in through the monitor daemon. Configuration edits preserve unrelated YAML, comments, and formatting, validate before writing, and serialize concurrent Phelix edits through a lock:

phelix.yamlyaml
security:
  vulnerability:
    enabled: true
    scan_interval: 24h
    fail_threshold: high

Exit codes: 0 below the configured fail_threshold, 25 at or above it, and 30 when OSV.dev is unreachable, including partial offline coverage with no findings, so a gate never passes on an unchecked project. See Exit codes.

Remote access and consent

The dashboard can view reports, dependency graphs, scan history, and advisory details, and can trigger scans or change settings, over the same authenticated gRPC channel as every other remote operation. Because dependency names and versions are private, every application opts in explicitly:

phelix.yamlyaml
security:
  vulnerability:
    reporting: true

With reporting disabled (the default), remote security operations against the application answer SECURITY_REPORTING_DISABLED and the monitor pushes nothing. With it enabled, the monitor pushes a summary of the latest completed scan on every stream sync, and the full reports, graphs, and configuration travel only on explicit backend queries. Remote mutations (running a scan, changing settings) additionally pass through the host authorization boundary (security action) and carry persistent idempotency, so a retried request never duplicates work.