Home /Examples/lib/examples/antivirus/MODERATE-interpreting-results.md

MODERATE-interpreting-results.md

lib/examples/antivirus/MODERATE-interpreting-results.md · 40 line(s) · markdown

Antivirus example (moderate): interpret results and coverage

A verdict is only useful if you read it correctly. This example covers the three things people most often get wrong about the integrity preflight.

1. PASS is scoped to the manifest

The check compares files listed in the manifest against their recorded digests. Files that are not in the manifest are simply not checked. So:

Know what your manifest covers before you rely on a PASS.

2. A mismatch names the file

When a covered file changed, the tool reports which one. Two honest responses:

3. An error is not a PASS

A tool error, a missing manifest, or an unavailable scanner is not a clean verdict. Never interpret "could not check" as "safe." Fix the error and re-run.

Try it

python3 tools/verify-before-execution.py \
  --manifest security/sha256-manifest.json \
  --root .

Expect a PASS on an unmodified checkout. Edit any covered file and re-run to see the mismatch report (then restore the file or regenerate the manifest).