agentleFS
michaelalbermichaelalber/ai-toolkit/skills/oss-vetting/SKILL.md

oss-vetting

Skill3 starsChanged 4 months agoMITView on GitHub

OSS library vetting and SBOM analysis for federal contractor environments (LANL/DOE/CUI). Use this skill when evaluating any open-source or third-party package for use in a federal contractor project — including NuGet packages, npm, PyPI, or any OSS library. Triggers include: vet this library, can we use X in LANL, is X approved for CUI systems, evaluate this NuGet package, generate an SBOM, supply chain risk, check this dependency, OSS assessment, 800-171 compliance for this package, C-SCRM review. Also trigger for any question about whether a library is safe, compliant, or appropriate for use on a government contract or CUI-adjacent system.

The file

We split this skill into the rules it gives an agent

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

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Exit criteria: every dimension scored with cited evidence.

    line 65

≈ “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

OSS Vetting Skill

"On a federal contract, a dependency is not a convenience — it is an attack surface you inherited and a compliance liability you signed for. Vet it like both."

Core Philosophy

Producing an OSS Vetting Assessment Report for federal contractor use is a compliance act, not a checkbox. Every assessment is evaluated against four frameworks — understanding why each matters prevents shallow checkbox compliance:

FrameworkRole
EO 14028Creates the legal obligation — software on federal systems must be verifiably secure
NIST SP 800-218 (SSDF)Defines the process — how maintainers should build and respond to vulnerabilities
NIST SP 800-171 Rev 2Sets the data-protection stakes — 110 controls for CUI systems; a vulnerable dep is a direct compliance liability
NIST SP 800-161 (C-SCRM)Governs provenance — where the software came from, who maintains it, whether the supply chain is auditable

Non-Negotiable Constraints:

  1. SCOPE FIRST — never assess without package, version, ecosystem, and target system (CUI? Yellow Network? air-gapped?). A vague request produces a shallow report; ask.
  2. EVIDENCE, NOT CLAIMS — every dimension score cites a finding (CVE ID, repo signal, license SPDX), never an impression.
  3. AIR-GAP SAFE — for Yellow Network artifacts, no live-connectivity URLs, no cloud scanners (Snyk/Dependabot), offline-capable SBOM tooling only.
  4. RED FLAGS ESCALATE — a disqualifying finding (see references/assessment-dimensions.md) halts approval pending waiver, regardless of other scores.

Plain-language framework explanations for stakeholder briefings live in references/framework-glossary.md.

Workflow

DETECT    Gather inputs: package, version, ecosystem; target system + classification;
          intended use; constraints. Missing context → ask. No report written yet.

RESEARCH  Investigate the four dimension groups — identity/provenance, security posture,
          development health (SSDF), supply chain (C-SCRM). Web search if available;
          note limitations if air-gapped. Detail: references/assessment-dimensions.md.

SCORE     Rate each dimension 1–5 (5 = lowest risk): security posture, supply chain
          integrity, maintainership health, license compatibility, CUI suitability.
          Each score cites evidence. Scale + dimensions: references/assessment-dimensions.md.

LICENSE   Resolve SPDX identifier → risk tier (acceptable / review / disqualifying) and
          note attribution + patent clauses. Full matrix: references/license-matrix.md.

SBOM      Note SBOM availability and pick offline-capable tooling for the deployment context.
          Tooling table + air-gap rules: references/assessment-dimensions.md.

REPORT    Write the Confluence-ready report from references/report-template.md with a
          clear recommendation: Approve / Approve with conditions / Reject.

Exit criteria: every dimension scored with cited evidence; license tier resolved; SBOM posture noted; a single explicit recommendation rendered; red flags surfaced and escalated if present.

State Block

<oss-vetting-state>
phase: DETECT | RESEARCH | SCORE | LICENSE | SBOM | REPORT | COMPLETE
package: [name@version]
ecosystem: nuget | pypi | npm | other
target_system: [name] (cui: true|false, yellow_network: true|false, air_gapped: true|false)
license_spdx: [id] | unknown
license_tier: acceptable | review | disqualifying | unresolved
dimension_scores: security=_ supply=_ maintainer=_ license=_ cui=_ (1–5, blank = not-yet-scored)
red_flags: [count] | none
recommendation: approve | approve-with-conditions | reject | pending
last_action: [description]
next_action: [description]
</oss-vetting-state>

Output Template

  • Full report — references/report-template.md (Confluence-ready: header + classification marking, executive summary, dimension scores, findings, license analysis, supply-chain/SBOM notes, mitigations, references).
  • License analysis — risk tiers, attribution, and patent notes from references/license-matrix.md.
  • Stakeholder framework briefings — plain-language explanations from references/framework-glossary.md.

Render as Confluence-ready Markdown: ## section headers, tables for scores and license data, inline code for package names / CVE IDs / SPDX identifiers, a blockquote for the executive-summary callout, and the classification marking as a bold header line at the top.

Integration with Other Skills

SkillRelationship
supply-chain-auditRun to scan the consuming project's full dependency tree once a package is approved.
dependency-mapperMap where the vetted package sits in the project's dependency graph and its blast radius.
security-review-federalPairs the package assessment with a federal security review of the consuming system.
nuget-package-scaffold / pypi-package-scaffoldApply vetting standards when authoring first-party packages, not just consuming third-party ones.

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.