agentleFS
Sign inStart free
narenaryannarenaryan/agent-skills/skills/security/rbvm-applicability/SKILL.md

rbvm-applicability

Skill0 starsChanged 39 days agoMITView on GitHub

Use when package matches conflict with backports, OSV revisions or withdrawals, scoped VEX assertions, or reachability evidence, especially when an old suppression survives changed advisory data.

The file

We split this skill into the rules it gives an agent

1 rule. About 0 are common, 0 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈0

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Run python3 skills/security/rbvm-applicability/reproduce.py for original offline revision/invalidation scenarios, not a scanner or …

    line 34

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

RBVM Applicability

Applicability is a versioned conclusion about a build and configuration, not a permanent property of a CVE. Its evidence dependencies can change without a deployment. Withdrawal invalidates a source record; it does not install a fix.

Interpret without flattening

EvidenceSemantics
OSV rangesintroduced inclusive; fixed exclusive; last_affected inclusive
OSV limitExclusive scope boundary, not proof of a fix
OSV identityaliases means equivalent vulnerability; upstream and related are different relationships
OSV revisionFor the same id, later modified wins; retrieval time is not revision order
OSV withdrawnEntry withdrawn at that time; retain history and reassess other evidence
Vendor backportCompare full distribution package release and advisory, not upstream version alone
OpenVEX timeStatement timestamp overrides document timestamp, then encapsulating timestamp
OpenVEX versionRevision within that document, not a ranking across publishers/documents

Recipe

  1. Resolve ecosystem, package instance/version, distribution/release, architecture, digest, and configuration. Preserve source bytes, issuer, IDs, revision metadata, collection time, and accepted authority scope.
  2. Reconcile revisions before matching. Compare OSV modified only for the same ID from accepted sources. Quarantine divergent equal-revision payloads and invalid/future timestamps under local clock policy. Keep withdrawal records so delayed older deliveries cannot resurrect them. A failed fetch or missing search result is not a withdrawal.
  3. Evaluate explicit versions OR applicable OSV ranges using ecosystem ordering; Git ranges require repository/commit ancestry. Unsupported comparisons remain unknown. Check exact vendor patch evidence: an older upstream version may contain a backport; possibility is not proof.
  4. Validate VEX authority, product/component/build scope, effective statement time, revision, and rationale. not_affected needs justification or impact statement; free-text-only reasoning needs review. Preserve untouched statements' original inherited time when updating documents. Do not let a newer unrelated document or publisher silently resolve contradictory claims. under_investigation cannot suppress.
  5. Bind each reviewed verdict to artifact/configuration and every relied-on advisory/assertion snapshot. Changed ranges, relationships, withdrawal, VEX status, trust policy, or scope invalidate dependent conclusions. Recompute from remaining evidence; do not erase alias siblings or promote unknown to clean. Missing/stale/conflicting inputs enter reassessment with an owner, not silent continued suppression.
  6. Save affected/not-affected/fixed/under-investigation disposition, reasons, reviewer, and next review trigger separately from enforcement. No evidence refresh grants permission to change suppressions, patch, deploy, or close tickets.

Counterexample: advisory A is withdrawn while alias B still identifies the installed package as affected. Retire A's claim, not the vulnerability globally. A newer accepted VEX under_investigation also defeats reuse of yesterday's not_affected decision; it proves neither affectedness nor a repair.

Run python3 skills/security/rbvm-applicability/reproduce.py for original offline revision/invalidation scenarios, not a scanner or production evaluator.

Sources: OSV v1.9.2, OpenVEX v0.2.0, Red Hat, reviewed 2026-10-24. Reassessment, trust and enforcement gates are local policy; SOURCES.md.

Pitfalls

  • SemVer ordering is invalid for arbitrary RPM/Debian versions.
  • A backport possibility is not evidence that this installed build is patched.
  • Alias deduplication must not erase per-package applicability or bundle unrelated fixes.
  • Runtime non-observation does not prove unreachable code.
  • A fresh download cannot rejuvenate old evidence or establish publisher authority.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free