Skip to content

What is TrustSight?

See beyond the package.

TrustSight is a security analysis and intelligence platform for the Arch User Repository ecosystem. It audits a PKGBUILD before you build it: it reads the diff and the post-patch recipe, applies published detection rules, and reports what changed, what matched, and what it could not see.

What it analyses

  • The change. A review compares the previous commit with the current one, so it can tell a version bump from a maintained package being quietly repointed.
  • The recipe as it will be read. Variables, arrays, function boundaries and source URLs are resolved where static resolution allows. The analysis never executes the PKGBUILD.
  • Ecosystem signals. Maintainer history, package age, adoption, release cadence, dependency relationships and cross-package correlation place the change in context.
  • Declared verification. Checksums, PGP keys and pinning are reported, not credited. A declaration an attacker can assert for free must not lower a score.

A platform, not only a diff check

TrustSight is a security analysis and intelligence platform for the Arch User Repository ecosystem. It combines static analysis, supply-chain intelligence, behavioral heuristics, dependency graph analysis, threat intelligence, and historical correlation to uncover suspicious package updates and emerging threats.

Beyond package analysis, TrustSight tracks ecosystem activity and lifecycle signals, including maintainer histories and changes, package adoptions and orphan status, ownership transitions, release patterns, and dependency relationships. It uses these signals to identify unusual behavior, suspicious takeovers, stale package revivals, and coordinated activity across packages.

It transforms package changes and ecosystem signals into deterministic, evidence-linked security assessments, helping you understand what changed, who is behind it, and what it might mean before you install.

What a run produces

A review is a deterministic, evidence-linked report:

  • Findings: a named rule with the file, side and line it matched, and a plain-English description.
  • A change summary: what the update altered, including source and checksum moves, dependency additions and maintainer or version changes.
  • Coverage gaps: what the run could not fully assess, stated outright so an incomplete analysis is never mistaken for a clean one.
  • A template verdict: the findings rendered into a review direction. It describes what fired; it never grants permission to skip review.

The default output is evidence, not a verdict, and nothing it prints is a promise of safety. The evidence is what you decide on. See How TrustSight Works for the full pipeline and Understanding Your First Report.

What it is not

TrustSight is an instrument, not a judge. It reports evidence, and the absence of evidence, and never substitutes its report for your decision. Absence of alerts is not a promise of safety. It does not decide whether a package is safe, and a quiet result is not a certificate: it means no monitored condition matched in the material examined. The security model states the boundaries behind that promise, and how each one is enforced.

The security model states the guarantees and their enforcement. The threat model states the limits.

Where to go next

  1. Install TrustSight.
  2. Run your first review.
  3. Understand the report.
  4. Pick a workflow: auditing before an update, gating yay and paru, or using TrustSight in CI.