go-security-audit — скилл для Claude Code: что делает и как установить

---
name: go-security-audit
description: >
  Security review for Go applications: input validation, SQL and command
  injection, SSRF, path traversal, auth, secrets, cryptography, TLS, CSRF and
  cookies, OWASP Top 10. Use when performing security reviews, checking for
  vulnerabilities, hardening Go services, or reviewing auth implementations.
  Trigger examples: "security review", "check vulnerabilities", "OWASP",
  "SQL injection", "input validation", "secrets management", "auth review".
  Not for: dependency CVEs (go-dependency-audit), concurrency safety
  (go-concurrency-review).
user-invocable: true
license: MIT
compatibility: Designed for Claude Code or similar AI coding agents working on Go projects. Requires the Go toolchain. govulncheck, gosec and gitleaks are optional. Read-only: this skill reports findings, it does not edit code.
allowed-tools: Read Glob Grep Bash(go:*) Bash(gofmt:*) Bash(govulncheck:*) Bash(gosec:*) Bash(gitleaks:*)
metadata:
  author: eduardo-sl
  version: "1.8.0"
---

# Go Security Audit

Security is not a feature — it's a property. Every line of code either
maintains it or degrades it.

## Operating Modes

Pick the mode that matches the request before starting:

- **Targeted check** — a single concern ("is this query injectable?",
  "review this auth middleware"). Apply only the relevant sections.
- **Diff audit** — audit the changed lines of a PR or working tree for
  every concern below.
- **Full audit** (default for "security review the service") — sweep the
  codebase using the parallel passes in "Auditing Large Codebases".

## Research Before Reporting

- Trace each suspicious sink back to an entry point (go-semantic-tools:
  references, call hierarchy). A sink fed only by constants or internal
  values is not a vulnerability.
- Check what upstream already enforces — type parsing, allowlists,
  middleware. It lowers severity; it does not delete the finding.
- 🔴 BLOCKER: exploitable today by whoever controls the input. 🟡 WARNING:
  needs a second failure or a privileged caller. 🟢 SUGGESTION: hardening.
- Every finding states file:line, the entry → sink path, a one-sentence
  exploit, and the fix.

Trust boundaries, STRIDE checks and the finding template:
`references/threat-model.md`.

## Run the Scanners First

Before manual review, run the automated scanners and fold their output
into the findings (skip any that is not installed and note it):

```bash
govulncheck ./...       # known CVEs actually reachable from your code
gosec ./...             # static analysis for insecure patterns
go vet ./...            # includes some security-relevant checks
```

Scanners find the known patterns; the manual passes below find the
logic flaws they cannot.

## Auditing Large Codebases

Each numbered section below is an independent audit pass. For codebases
beyond ~20 files:

1. Locate the attack surface first: HTTP/gRPC handlers, CLI entry
   points, queue consumers, and anything parsing external input.
2. Run one pass per concern: (a) input validation + SQL and other
   injection (command, SSRF, paths, archives, templates),
   (b) authentication/authorization, (c) secrets + crypto,
   (d) TLS + security headers + rate limiting, (e) logging hygiene.
3. If your environment supports delegating work to parallel sub-agents
   or tasks, assign each pass to one — the passes don't overlap.
   Otherwise run them sequentially.
4. Every finding must cite `file.go:line`, the vulnerable input path,
   and a concrete fix. Aggregate into one report sorted by severity.

Detailed reference material, loaded on demand:

- `references/injection.md` — boundary validation, sanitization,
  parameterized and dynamic queries, command injection, SSRF, path
  traversal, archives, templates, open redirects, log injection.
- `references/auth-and-transport.md` — passwords, JWT, authorization
  middleware, secrets, headers, cookies, CSRF, CORS, TLS, rate limiting.
- `references/threat-model.md` — trust boundaries, STRIDE per boundary,
  severity definitions, finding template.
- `references/crypto.md` — randomness, constant-time comparison, password
  hashing, AEAD encryption, hashes and KDFs, signatures, FIPS mode.

Read a reference file only when the section below is not enough.

## 1. Input Validation

Validate at the boundary, before the value reaches any business code:

- Cap the request body with `http.MaxBytesReader` — an unbounded decoder is
  a memory exhaustion vector.
- Decode into a typed struct, then validate it. A decode that succeeds is not
  a value that is valid.
- Reject rather than repair. Reply with the status code, not with the
  internal error text.
- Sanitize anything rendered back as HTML (`bluemonday`), parse emails with
  `net/mail`, and reject URLs whose scheme is not `http`/`https`.

Examples in `references/injection.md`.

## 2. SQL Injection Prevention

ALWAYS pass values as query parameters. There is no safe amount of string
concatenation:

```go
// ✅ Good — parameterized
row := db.QueryRowContext(ctx, "SELECT id, name FROM users WHERE email = $1", email)

// ❌ CRITICAL — SQL injection
query := "SELECT * FROM users WHERE email = '" + email + "'"
query := fmt.Sprintf("SELECT * FROM users WHERE id = %s", id)
```

Dynamic filters are built by appending placeholders (`$1`, `$2`, …) and
collecting the values into an `args` slice — never by interpolating the value
itself. Table and column names cannot be parameterized: allowlist them
against a fixed set.

Examples in `references/injection.md`.

## 3. Injection Beyond SQL

- **Command** — `exec.CommandContext(ctx, "tool", "--", arg)`: no shell,
  `--` before input. Input never selects the binary.
- **SSRF** — server-side fetches of caller URLs go through a client whose
  dialer rejects loopback, private, link-local and metadata addresses
  *after* DNS resolution, with proxies disabled.
- **Paths** — `os.Root` (Go 1.24+) for any user-influenced path;
  `filepath.Join` + `HasPrefix` misses symlinks.
- **Archives** — reject non-local entry names (zip slip), cap entry count,
  declared size and bytes actually read (decompression bombs).
- **HTML** — `html/template` only; `template.HTML(input)` disables escaping.
- **Redirects** — only same-site relative paths or allowlisted hosts.

Examples in `references/injection.md`.

## 4. Authentication & Authorization

- Hash passwords with `bcrypt` (or argon2id). NEVER store plaintext, NEVER
  use MD5/SHA — they are fast, which is the wrong property here.
- Compare with `bcrypt.CompareHashAndPassword`, which is constant-time.
  Hand-rolled comparisons leak timing.
- Validate every JWT claim that matters: signature with the expected
  algorithm, `exp`, `iss`, `aud`. Reject `alg: none` and never hardcode the
  signing key.
- Authorize per request in middleware, reading the identity from the request
  context. An endpoint with no role check is a public endpoint.
- Check ownership, not only role: a valid token for user A must not read
  user B's row.

Examples in `references/auth-and-transport.md`.

## 5. Secrets Management

- 🔴 NEVER hardcode secrets, tokens, or API keys in source code
- 🔴 NEVER commit secrets to git (even in "test" files)
- 🔴 NEVER log secrets, tokens, or passwords

Read them from the environment or a secrets manager, keep `.env`, `*.pem`,
`*.key` and `credentials.json` in `.gitignore`, and run `gitleaks detect` in
CI so a leak fails the build instead of ageing in history.

Examples in `references/auth-and-transport.md`.

## 6. Cryptography and Randomness

- `crypto/rand` for anything that must not be guessed — tokens (`rand.Text`,
  Go 1.24+), nonces, salts. `math/rand` never.
- Compare secrets with `subtle.ConstantTimeCompare` / `hmac.Equal`.
- Encrypt with AES-GCM via `cipher.NewGCMWithRandomNonce` (Go 1.24+), so
  no nonce is ever reused. No ECB, no unauthenticated modes.
- SHA-256/SHA-3 for integrity, HMAC for authentication, HKDF to derive keys,
  Ed25519 for signatures. MD5 and SHA-1 never protect anything.
- FIPS environments build with `GOFIPS140` (Go 1.24+).

Examples in `references/crypto.md`.

## 7. Transport, Headers and Rate Limiting

- Set `X-Content-Type-Options: nosniff`, `X-Frame-Options: DENY`, a
  `Content-Security-Policy`, and HSTS on every response.
- TLS: raise `MinVersion` to `tls.VersionTLS12` (or 1.3) and keep Go's
  default cipher suites — a hand-picked list only goes stale.
  NEVER `InsecureSkipVerify: true` outside a test.
- Rate limit auth endpoints, public APIs, and anything expensive, keyed per
  client rather than globally.
- Browser-facing: session cookies `Secure`, `HttpOnly`, `SameSite`, with a
  `__Host-` prefix; CSRF via `http.CrossOriginProtection` (Go 1.25+), with
  no state change on GET; CORS from an exact-origin allowlist, never `*`
  on authenticated APIs.
- pprof and other debug handlers only on an internal listener.

Examples in `references/auth-and-transport.md`.

## 8. Logging Security

Log the identifier, never the credential:

```go
// ❌ CRITICAL
log.Printf("user login: email=%s password=%s", email, password)
log.Printf("request body: %v", req) // may contain secrets

// ✅ Good — redacted
logger.Info("auth completed", slog.String("user_id", userID))
```

Whole-struct logging is the common leak: a `%v` on a request struct prints
whatever field someone adds next sprint.

## Security Audit Checklist

### Critical (🔴 BLOCKER)
- No SQL injection vectors (all queries parameterized)
- No shell invocation with input; no user-controlled path outside an `os.Root`
- Server-side fetches of user URLs blocked from private/metadata addresses
- No hardcoded secrets/keys/tokens
- No plaintext password storage
- No `math/rand` for tokens, IDs or nonces; no `==` on secrets
- No disabled TLS certificate verification
- Request body size limited
- JWT signature verified, `alg: none` rejected

### Important (🟡 WARNING)
- Input validation on all external data
- Rate limiting on auth and public endpoints
- Security headers set on all responses
- CORS configured restrictively
- Error messages don't leak internals
- Audit logging for auth events

### Recommended (🟢 SUGGESTION)
- `govulncheck` in CI pipeline
- `gitleaks` for secret scanning
- Structured logging with redaction
- Dependency pinning with verified checksums