Your portable security pre-flight kit. Run it before you work in any repo.

Run it

git clone https://github.com/segraef/sec-kit.git && cd sec-kit
bash seckit.sh              # macOS / Linux
pwsh ./seckit.ps1           # Windows

Action	What it does
doctor	Reports which scanners and clients (jq, yq, gh, az, gitleaks, semgrep, checkov, osv-scanner, trufflehog, pre-commit) are installed and which are missing.
install	Installs every missing scanner and client via brew/npm/pipx/scoop. Run this once on a fresh machine.
scan	Sweeps a folder of repos for vulnerable dependencies, code/IaC flaws, malware and secrets. Pick all scanners or a subset (osv, gitleaks, trufflehog, semgrep, checkov, socket).
scan-skill	Statically vets an AI agent skill or MCP server (directory, .zip or git URL) before you install it: prompt-injection, data exfiltration, credential theft, supply-chain RCE, obfuscation, over-broad agency and MCP tool poisoning. Never executes the target; prints a 0-100 risk verdict and a markdown report.
harden	Drops pre-commit, gitleaks, SECURITY.md, CODEOWNERS, dependabot, CodeQL and PR templates into a repo so the next commit is clean. On Node repos it also sets ignore-scripts=true in .npmrc to block install-time supply-chain worms (see below).
agent	Installs the SecKit prompt as a Claude subagent, Copilot chat mode, Cursor rule or AGENTS.md section so any AI assistant runs the same playbook.
mcp	Wires the official MCP servers (Semgrep, Snyk, OSV, Trivy, Scorecard, GitHub, ADO, Atlassian, Microsoft Learn, Terraform, Foundry) into Claude/Copilot/Cursor.
audit	Read-only posture check against a GitHub org/repo or Azure DevOps project/repo. Safe to run anywhere because every call is a GET.
enforce	Writes the missing settings flagged by audit. Dry-run by default; pass --apply / -Apply to actually write.
reminders	Prints every security reminder in the kit. Handy as a checklist.

Run it in CI

Drop-in pipelines that run the same flow on every push: seckit install provisions the scanners, seckit scan sweeps the repo, and the markdown report is published as a build artifact. They clone SecKit at run time, so the only thing your repo needs is the one file.

• GitHub Actions: .github/workflows/seckit-scan.yml
• Azure Pipelines: .pipelines/seckit-scan.yml

Both are soft-fail by default (findings produce a warning plus the report artifact, not a red build); flip the gate step to exit 1 / remove continueOnError to block merges on findings.

Blocking npm install-script worms

Self-propagating npm worms (for example the May 2026 redhat-cloud-services worm that hit 90+ packages) run their payload from a package's preinstall/install/postinstall hook during npm install, before any of your own code runs. A scanner only helps if it runs before install, and only catches what it has catalogued; by the time a brand-new variant has an advisory, the hook has already executed and exfiltrated your npm/GitHub/cloud/SSH tokens.

seckit harden closes the vector at the source on any repo with a package.json: it appends ignore-scripts=true to .npmrc, so no dependency lifecycle script executes on install. This holds even for a variant no scanner knows about yet.

The trade-off, handled. A few legitimate deps build native code in those hooks (esbuild, sharp, bcrypt, ...), and your own root postinstall/prepare (husky, prisma) is skipped too. harden does not leave you to discover this the hard way: it scans node_modules and prints the exact deps in your repo that build via install scripts, and the generated .npmrc documents how to allowlist them rather than disabling the protection:

npx --yes @lavamoat/allow-scripts auto   # write a vetted allowlist into package.json
npx --yes @lavamoat/allow-scripts        # run ONLY allowlisted scripts, after install
# or, dependency-free, for a single vetted native dep:
npm rebuild <pkg> --ignore-scripts=false

A hardened install flow then looks like npm ci && npx --yes @lavamoat/allow-scripts && npm run prepare.

Releases

Releases are automated with release-please, driven by Conventional Commits on main:

• Push a feat:/fix: commit to main. The release workflow opens (or updates) a release PR that bumps version.txt + .release-please-manifest.json and rewrites CHANGELOG.md from your commits.
• Merge that PR. release-please tags the commit vX.Y.Z and publishes a GitHub Release.

Pre-1.0, feat: bumps the minor and fix: bumps the patch (configured in release-please-config.json). seckit version prints the current version. One-time repo setting: Settings > Actions > General > "Allow GitHub Actions to create and approve pull requests" must be enabled so the release PR can be opened with the default token.

More: docs/ · CONTRIBUTING.md · CHANGELOG.md · LICENSE