vet: Skill for Claude Code

.claude/skills/vet-policy-authoring/SKILL.md

vet-policy-authoring is a skill for Claude Code from safedep/vet. It costs 67 tokens per session (874 once invoked), scanned A, original, Apache-2.0.

A guide for writing and testing a vet policy, which is a YAML file containing rules that fail or warn on scan findings and optional suppressions. It also explains how vet reads and applies that policy.

In plain words
What is it for?
Use it to draft a new policy, test it against a saved scan, add a warning or failure rule, or suppress a finding with a reviewable reason.
Why use it?
It helps create a policy with valid fields and check its behavior without silently installing it or changing project configuration.

Skill for Claude Code

Written for Claude Code: installed under .claude/.

This is safedep/vet's own configuration. It tells Claude Code how to work on vet itself, so it is not a mod to install elsewhere. Copy it as a starting point and replace the rules that are about this project. Everything vet configures →

About the project

vet is a software supply-chain security tool that examines open-source packages for malware, vulnerabilities, license issues, and policy violations. Developers use it to scan package manifests, repositories, container images, or SBOMs and to enforce security policies in CI/CD systems.

safedep/vet · 1,113 stars · on GitHub · safedep.io

Reuse

Borrowing it

Nothing to install: this file belongs to safedep/vet. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file
curl -O https://raw.githubusercontent.com/safedep/vet/v2/.claude/skills/vet-policy-authoring/SKILL.md
Clone the repo
git clone --depth 1 https://github.com/safedep/vet

Made for: Claude Code.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for vet-policy-authoring

README.md
[![agentmods](https://agentmods.dev/badge/skills/safedep/vet/vet-policy-authoring/github.svg)](https://agentmods.dev/skills/safedep/vet/vet-policy-authoring)
Your own site
<a href="https://agentmods.dev/skills/safedep/vet/vet-policy-authoring"><img src="https://agentmods.dev/badge/skills/safedep/vet/vet-policy-authoring/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for vet-policy-authoring

Your own site · 80×15
<a href="https://agentmods.dev/skills/safedep/vet/vet-policy-authoring"><img src="https://agentmods.dev/badge/skills/safedep/vet/vet-policy-authoring.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 67 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 874 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00067 $0.00874
Opus 5.5 $0.00027 $0.00350
Sonnet 5.5 $0.00013 $0.00175
Haiku 4.5 $0.00007 $0.00087

Measured 4d ago against content hash 769bccefbf3e, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

vet-policy-authoring scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 4d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

.claude/skills/vet-policy-authoring/SKILL.md · 53 lines

How it starts

The opening of the file, as written. The whole thing — 53 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Write and test a vet policy

A vet policy is a YAML file with version: 2, a list of rules and a list of suppressions. A rule has an id, a CEL condition in when and an action, fail or warn. A suppression names an id, a purl or a control, a reason, and an optional expires date. A path suppression also names a control. vet reads a policy only with --policy FILE or the policy.file config key.

Rules for you

  • Draft the policy in a new file that the user names, for example vet-policy.yml. Do not change a policy file that exists unless the user asks.
  • Never install the policy. Do not set policy.file, and do not edit a CI workflow, a git hook or the vet config. Tell the user the command that applies the policy, and let the user run it.
  • Read the fields from the schema. Do not guess a field name or a control id.
  • Every suppression has a reason that a reviewer can check. Give it an expires date when the user knows when the fix comes.

Steps

  1. Read the input of a rule: vet policy schema get. A rule reads finding, package and manifest. An optional field with no value is absent, so test it with has(), for example has(package.days_since_publish) && package.days_since_publish < 5. Compare a package name with package.is("name"), not with ==, and a version with package.version_cmp("1.2.3"). Both apply the rule of the ecosystem: package.is("python-dateutil") matches python.dateutil, and package.version_cmp("2.0") < 0 holds for 1.0rc1. An ecosystem with no version order, such as GitHub Actions, gives no answer to package.version_cmp. A condition that needs the answer does not match, but CEL still decides a || b when b is true.
  2. Read the control ids and their default severities: vet policy control list -o json.
  3. Start from the starter file when the user has no policy: vet policy init vet-policy.yml.
  4. Write the rules. Keep one idea in each rule, and give it a description.
  5. Check the file: vet policy validate vet-policy.yml -o json. Fix every problem that it lists.
  6. Test the policy against a saved scan. vet report list -o json lists the scans. With no scan, run vet scan . -o json once. Then run vet report show last --policy vet-policy.yml -o json. The trailer.gate object holds the outcome, the rules that failed and the finding ids. Each finding that a rule or --fail-on matched has a gate object: the action (fail or warn), the rules, the fail_on value and the help and link of the first rule. The saved scan does not change.
  7. Show the user the policy, the gate outcome, and the findings that each rule matched or that a suppression hid. Then give the command that applies it: vet scan . --policy vet-policy.yml.

Read the full file on GitHub · 53 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 4d ago Changed 769bccefbf3e
  2. 7d ago First seen · 53 lines · 67 tokens per session scan A 8fded3798581

Subscribe to this mod's changes

vet-policy-authoring is a skill published in the GitHub repository safedep/vet (1,113 stars, last pushed today), licensed Apache-2.0. It adds 67 tokens to every session and 874 once invoked, about $0.0003 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-10-10.

Related

Other skills, from other repositories

pr-draft-summary

Draft a concise pull request handoff after substantive repository changes are finished or ready for review. Trigger when wrapping up code, test, script, workflow, release, security, documentation-with-behavior-impact, or template customization changes and the user needs a PR title/body grounded in the real diff…

CaseyLabs/kc-secure-repo-template · 93 tokens

workflow-validation

Use when validating repository workflows after changes to Docker-first Makefile targets, Dockerfiles, scripts, CI, release packaging, smoke tests, or documentation alignment. Do not use for repo redesign, GitHub-policy-only changes, or instruction-only skill edits with no workflow impact.

CaseyLabs/kc-secure-repo-template · 57 tokens

supply-chain-security-scanner

AI-powered software supply chain security auditing skill for agentic platforms. Performs comprehensive dependency vulnerability scanning across npm, PyPI, Maven, Go modules, Cargo, and container images. Generates SBOMs (Software Bill of Materials) in SPDX and CycloneDX formats using Syft and Grype. Validates license…

JPeetz/agent-skills · 307 tokens

golang-testing

Go testing best practices including table-driven tests, test helpers, benchmarking, race detection, coverage analysis, and integration testing patterns. Use when writing or improving Go tests.

affaan-m/ECC · 37 tokens

ci

Configure Ginkgo for continuous integration — the recommended CLI flag set and the rationale for each flag (-r -p --randomize-all --randomize-suites --fail-on-pending --fail-on-empty --keep-going --cover --race --trace --json-report --timeout --poll-progress-after/-interval), invoking via go run to pin the CLI to…

onsi/ginkgo · 151 tokens

debugging-failures

Diagnose a failing Ginkgo suite as an agent — always run with --json-report into a predictable temp/gitignored location, read the terminal verdict line, then use jq to extract structured failure details (name, message, file:line, panic value, captured logs). Covers the panicked-vs-failed trap, panic locations pointing…

onsi/ginkgo · 122 tokens