Security Scan

Security scan — orchestrate scanners, triage the noise, emit a gate

A machine-first vulnerability sweep. Point automated scanners at a codebase, collect SARIF/JSON, then do the work that has actual value: dedupe cross-tool overlap, rank by exploitability, and emit one gate artifact CI can act on. The finding comes from a tool run, not a hunch — if you are reasoning about a design or hand-writing a fix, that is secure-coding, not this skill.

Your job is orchestration + triage: every finding traces to a scanner run with a ruleId and a source location, so the output is reproducible. Not eyeballing code, not authoring patches.

Read-only by default. Scan, triage, report. Apply fixes (version bumps, rotation, .gitignore edits) only when the user asks — a security sweep that silently mutates the tree destroys the evidence and the trust.

Pin and verify your scanners. Exact versions, verified checksums/SHAs, never @latest GitHub Actions. In March 2026 Trivy was supply-chain compromised — malicious releases v0.69.4/0.69.5/0.69.6 and a hijacked aquasecurity/trivy-action exfiltrated CI secrets. Your scanner runs with repo + CI-secret access; an unpinned scanner is itself the attack surface.

Scan classes and tool selection

Four classes: SAST (injection, XSS, path traversal in first-party code), SCA / deps (known CVEs in dependency manifests + lockfiles), Secrets (credentials in the tree or git history), Misconfig / IaC (Dockerfile, k8s, Terraform, exposed config). Pick tools by what is in the repo. This is the real branch point — match the tool to the manifest, do not run everything everywhere.

Full install (pinned), flag matrix, and suppression syntax: references/tools.md.

Run recipes

All recipes emit SARIF (or JSON you normalize to it) — a common schema is what lets you merge four tools, dedupe, and feed one artifact into CI instead of four incompatible logs. Pin the version shown; the placeholders below mark where to lock an exact tag/digest.

SAST — Semgrep

Free OSS edition (latest 1.164.0, 2026-05-27): 30+ languages, ~2,000 community rules. SCA + Secrets rulesets are gated behind the hosted platform — use the dedicated tools below for those, not Semgrep.

# Pin via the CLI version, not @latest. OWASP ruleset, SARIF out.
semgrep scan --config p/owasp-top-ten --sarif --output sast.sarif .
# Broader local sweep (community rules), no telemetry:
semgrep scan --config auto --sarif --output sast.sarif --metrics off .

SCA — osv-scanner (primary), native auditors (fast pass)

osv-scanner (OpenSSF/Google) checks lockfiles against OSV.dev across ecosystems and catches transitive CVEs the native auditors miss. Run native first for speed, osv-scanner for coverage — never native alone.

# Primary: lockfile-aware, multi-ecosystem, SARIF.
osv-scanner scan source --format sarif --output sca.sarif .
# Fast first pass (ecosystem-native, weaker on transitive):
npm audit --omit=dev --audit-level=high --json > npm-audit.json   # Node
pip-audit --format json --output pip-audit.json                   # Python

Secrets — gitleaks (tree + history), TruffleHog (verified)

gitleaks (~150+ patterns, sub-second on diffs) is the pre-commit/CI workhorse. TruffleHog (800+ types) adds live credential verification — it auth-tests a hit to tell a real leaked key from a sample. Scan history, not just the tree: a key deleted in HEAD is still in the pack files and still rotatable.

# gitleaks: redacted SARIF over the working tree AND full git history.
gitleaks detect --redact --report-format sarif --report-path secrets.sarif
# TruffleHog: only verified (live) secrets across history.
trufflehog git file://. --only-verified --json > trufflehog.json

Misconfig / IaC — Trivy (PINNED — see the caveat above)

Trivy scans filesystems, images, and IaC and finds transitive CVEs npm audit misses.

# Pin the EXACT version (NOT v0.69.4/.5/.6 — those were the malicious releases).
# Verify the checksum/cosign signature before first use. See references/tools.md.
trivy fs --scanners vuln,secret,misconfig --format sarif --output trivy.sarif .

Triage — turn noise into a ranked report

The deliverable is not the four SARIF files. It is a deduped, ranked report.

1. Merge + dedupe. Cross-tool overlap is real (osv-scanner and Trivy both flag the same CVE; gitleaks and TruffleHog both flag the same key). Key on (class, normalized-id, path, line) and keep the richest record — prefer the one with verification (TruffleHog) or reachability (govulncheck).
2. Normalize severity. SARIF level and tool-native severities disagree; map them all to one critical/high/medium/low scale (references/triage.md).
3. Rank by exploitability, not by count. A finding scores higher when it is reachable (called, not just present) + exposed (on an untrusted path) + a sensitive sink (auth, money, PII, RCE). A verified live secret or a reachable RCE CVE outranks a theoretical lib finding behind a feature flag.
4. Suppress with a written justification, never a blanket ignore. Each suppression records who, why, and an expiry — not a silent .semgrepignore that hides the next real bug too.

BAD  — dump 412 raw findings from four tools, sorted alphabetically, no ranking.
GOOD — 3 unsuppressed criticals first:
       1. [secrets] VERIFIED live Stripe sk_live_… in config/.env (history) → ROTATE NOW
       2. [sca] CVE-2024-… lodash 4.17.20 transitive, reachable in src/api/parse.ts → bump 4.17.21
       3. [sast] SQL built from req.query in routes/search.js:48 → parameterize
       + 7 mediums summarized, + 18 suppressed (each with justification + expiry).

Full ranking rubric, dedupe keying, and severity-normalization map: references/triage.md.