proofctl 0.2.0__tar.gz → 0.3.1__tar.gz

{proofctl-0.2.0 → proofctl-0.3.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: proofctl
-Version: 0.2.0
+Version: 0.3.1
 Summary: Zero-dependency linter for Python, Terraform, Dockerfiles, Kubernetes, and GitHub Actions — catches AI slop and security misconfigurations pre-commit
 Author-email: Kolawolu Odunola <kolawolu.o@gmail.com>
 License: MIT

{proofctl-0.2.0 → proofctl-0.3.1}/proofctl/checkers/dockerfile.py

@@ -124,7 +124,7 @@ def _df001(instructions: list[tuple[int, str, str]], path: Path) -> list[Finding
             msg = f"Base image '{image}' uses the mutable ':latest' tag"
             hint = "Replace :latest with a pinned version and digest."
         else:
-            sev = Severity.WARNING
+            sev = Severity.INFO
             msg = f"Base image '{image}' is pinned to a tag but not a digest (still mutable)"
             hint = "Add digest pinning: FROM python:3.12@sha256:<digest>"
 

{proofctl-0.2.0 → proofctl-0.3.1}/proofctl/checkers/hcl_utils.py

@@ -3,12 +3,14 @@ from __future__ import annotations
 import re
 from dataclasses import dataclass, field
 
-# Matches the header of a block: keyword ("label1")? ("label2")? {
+# Matches the header of a block: keyword (label1)? (label2)? {
+# Labels may be quoted ("aws_vpc") or bare (aws_vpc). Terraform tolerates the
+# bare form on `resource <type> "name"`, e.g. terragoat's `resource aws_vpc`.
 _BLOCK_HEADER_RE = re.compile(
-    r'^\s*(\w+)'              # kind
-    r'(?:\s+"([^"]*)")?'      # optional label1
-    r'(?:\s+"([^"]*)")?'      # optional label2
-    r'\s*\{'                  # opening brace (rest of line may follow)
+    r'^\s*(\w+)'                            # kind
+    r'(?:\s+(?:"([^"]*)"|([A-Za-z_]\w*)))?' # optional label1: quoted or bare ident
+    r'(?:\s+(?:"([^"]*)"|([A-Za-z_]\w*)))?' # optional label2: quoted or bare ident
+    r'\s*\{'                                # opening brace
 )
 
 _HEREDOC_START_RE = re.compile(r'<<[-~]?(\w+)\s*$')
@@ -172,8 +174,10 @@ def _parse(lines: list[str], base: int, skip_outer: bool) -> list[HclBlock]:
             m = _BLOCK_HEADER_RE.match(line)
             if m:
                 block_kind = m.group(1)
-                block_label1 = m.group(2) or ""
-                block_label2 = m.group(3) or ""
+                # group 2 = quoted label1, group 3 = bare label1
+                block_label1 = m.group(2) or m.group(3) or ""
+                # group 4 = quoted label2, group 5 = bare label2
+                block_label2 = m.group(4) or m.group(5) or ""
                 block_start_idx = i
                 depth += delta
                 if depth == 0:

proofctl-0.3.1/proofctl/checkers/secrets.py

@@ -0,0 +1,446 @@
+"""Secret detection — PROOFCTL-SECRET-* rules. Scans any text file."""
+from __future__ import annotations
+
+import math
+import re
+from pathlib import Path
+
+from ..models import Finding, Severity
+from .base import FileChecker
+
+# ── allowlists ────────────────────────────────────────────────────────────────
+
+_SKIP_SUFFIXES = (
+    ".lock", ".min.js", ".bundle.js", ".map",
+)
+_SKIP_NAMES = frozenset({
+    "package-lock.json", "yarn.lock", "Cargo.lock", "go.sum", "pnpm-lock.yaml",
+    "composer.lock", "Pipfile.lock", "poetry.lock",
+})
+
+_PLACEHOLDER_VALUES = frozenset({
+    "changeme", "change-me", "change_me", "your-password-here",
+    "your_password_here", "yourpassword", "password", "passw0rd",
+    "xxx", "xxxx", "xxxxx", "xxxxxx", "placeholder", "todo", "tbd",
+    "example", "sample", "secret", "your-secret-here", "redacted",
+    "default", "test", "testing", "dummy", "fake", "none", "null",
+    "<password>", "<secret>", "<value>", "n/a", "na",
+})
+
+# ── patterns ──────────────────────────────────────────────────────────────────
+
+_PAT_AWS_KEY = re.compile(r"AKIA[A-Z0-9]{16}")
+_PAT_AWS_SECRET = re.compile(
+    r"""aws_secret(_access)?_key\s*=\s*["']([A-Za-z0-9/+=]{40})["']""",
+    re.IGNORECASE,
+)
+_PAT_GITHUB = re.compile(r"gh[pousr]_[A-Za-z0-9]{36,}")
+_PAT_SLACK = re.compile(r"xox[bpoars]-[A-Za-z0-9-]+")
+_PAT_PRIVKEY = re.compile(r"-----BEGIN [A-Z ]*PRIVATE KEY-----")
+_PAT_JWT = re.compile(r"eyJ[A-Za-z0-9_-]+\.eyJ[A-Za-z0-9_-]+\.[A-Za-z0-9_-]{20,}")
+_PAT_GENERIC = re.compile(
+    r"""(?P<key>password|passwd|pwd|secret|api[_-]?key|access[_-]?token|auth[_-]?token)"""
+    r"""\s*[:=]\s*["'](?P<val>[^"'$\{]{8,})["']""",
+    re.IGNORECASE,
+)
+_PAT_GOOGLE = re.compile(r"AIza[0-9A-Za-z_-]{35}")
+_PAT_STRIPE = re.compile(r"(?:sk|pk)_(?:test|live)_[0-9a-zA-Z]{24,}")
+_PAT_TF_ADMIN = re.compile(
+    r"""(?P<key>administrator_login_password|admin_password|root_password)"""
+    r"""\s*=\s*"(?P<val>[^"$\{][^"]{4,})\"""",
+)
+
+_ENV_REF_RE = re.compile(r"^\s*[\$\{][^\s]+[\}]?\s*$|^\$[A-Z_][A-Z0-9_]*$|^\$\{[^}]+\}$")
+
+# SECRET-011: kind: Secret YAML detection
+_PAT_K8S_KIND_SECRET = re.compile(r'^\s*kind\s*:\s*["\']?Secret["\']?\s*(#.*)?$')
+_PAT_YAML_DOC_SEP = re.compile(r'^---\s*(#.*)?$')
+_PAT_K8S_NAME = re.compile(r'^\s{2,}name\s*:\s*["\']?([A-Za-z0-9._-]+)["\']?\s*(#.*)?$')
+_PAT_K8S_DATA_KEY = re.compile(r'^(\s*)(data|stringData)\s*:\s*(#.*)?$')
+# Match indented "key: value" lines under data:/stringData:
+_PAT_K8S_DATA_ENTRY = re.compile(
+    r'^(\s+)([A-Za-z0-9._-]+)\s*:\s*(.*?)\s*(?:#.*)?$'
+)
+_K8S_SKIP_PATH_TOKENS = ("/tests/", "/fixtures/", "/testdata/")
+
+# SECRET-012: heredoc tracking + secret-shaped env assignment inside heredoc
+_PAT_HEREDOC_OPEN = re.compile(r'<<[-~]?(\w+)\s*$')
+_PAT_HEREDOC_SECRET = re.compile(
+    r'(?:export\s+)?([A-Z][A-Z0-9_]*(?:_KEY|_TOKEN|_SECRET|_PASSWORD|_PASSWD))'
+    r'\s*=\s*([^\s"\']+|"[^"]+"|\'[^\']+\')'
+)
+
+
+def _shannon_entropy(s: str) -> float:
+    if not s:
+        return 0.0
+    freq: dict[str, int] = {}
+    for ch in s:
+        freq[ch] = freq.get(ch, 0) + 1
+    n = len(s)
+    return -sum((c / n) * math.log2(c / n) for c in freq.values())
+
+
+def _is_placeholder(val: str) -> bool:
+    low = val.lower().strip()
+    if low in _PLACEHOLDER_VALUES:
+        return True
+    # Entirely lowercase + dashes/underscores looks like placeholder text
+    if re.fullmatch(r"[a-z][a-z0-9_\-]*", val) and any(
+        tok in low for tok in ("change", "your", "placeholder", "example", "sample")
+    ):
+        return True
+    return False
+
+
+def _is_env_ref(val: str) -> bool:
+    return bool(_ENV_REF_RE.match(val)) or "${" in val or val.startswith("$(")
+
+
+def _is_binary(data: bytes) -> bool:
+    if not data:
+        return False
+    if b"\x00" in data:
+        return True
+    text_chars = bytes(range(32, 127)) + b"\n\r\t\f\b"
+    nonprint = sum(1 for b in data if b not in text_chars)
+    return (nonprint / len(data)) > 0.30
+
+
+def _should_skip(path: Path) -> bool:
+    name = path.name
+    if name in _SKIP_NAMES:
+        return True
+    for suf in _SKIP_SUFFIXES:
+        if name.endswith(suf):
+            return True
+    # *.lock.json
+    if name.endswith(".lock.json"):
+        return True
+    return False
+
+
+# ── rule fns ──────────────────────────────────────────────────────────────────
+
+def _mk(path: Path, line: int, rid: str, name: str, sev: Severity, msg: str) -> Finding:
+    return Finding(
+        file=str(path),
+        line=line,
+        col=0,
+        rule_id=rid,
+        rule_name=name,
+        severity=sev,
+        message=msg,
+        hint="Move secret to env var, secret manager, or runtime injection. Never commit credentials.",
+        authority="OWASP ASVS V2 / CWE-798 – Use of Hard-coded Credentials",
+        docs_url="https://cwe.mitre.org/data/definitions/798.html",
+    )
+
+
+def _scan_line(path: Path, lineno: int, line: str) -> list[Finding]:
+    out: list[Finding] = []
+    # Track (rule_priority, secret_value) so most-specific rule wins per value.
+    # Keyed by value text → (priority, Finding)
+    by_value: dict[str, tuple[int, Finding]] = {}
+
+    def _claim(value: str, priority: int, finding: Finding) -> None:
+        existing = by_value.get(value)
+        if existing is None or priority < existing[0]:
+            by_value[value] = (priority, finding)
+
+    # SECRET-001: AWS access key
+    for m in _PAT_AWS_KEY.finditer(line):
+        v = m.group(0)
+        _claim(v, 1, _mk(path, lineno, "PROOFCTL-SECRET-001",
+                         "AWS access key ID", Severity.ERROR,
+                         f"Hardcoded AWS access key ID detected ({v[:8]}…)"))
+
+    # SECRET-002: AWS secret key
+    for m in _PAT_AWS_SECRET.finditer(line):
+        v = m.group(2)
+        _claim(v, 2, _mk(path, lineno, "PROOFCTL-SECRET-002",
+                         "AWS secret access key", Severity.ERROR,
+                         "Hardcoded AWS secret access key detected"))
+
+    # SECRET-003: GitHub token
+    for m in _PAT_GITHUB.finditer(line):
+        v = m.group(0)
+        _claim(v, 3, _mk(path, lineno, "PROOFCTL-SECRET-003",
+                         "GitHub token", Severity.ERROR,
+                         f"Hardcoded GitHub token detected ({v[:7]}…)"))
+
+    # SECRET-004: Slack token
+    for m in _PAT_SLACK.finditer(line):
+        v = m.group(0)
+        _claim(v, 4, _mk(path, lineno, "PROOFCTL-SECRET-004",
+                         "Slack token", Severity.ERROR,
+                         f"Hardcoded Slack token detected ({v[:6]}…)"))
+
+    # SECRET-005: Private key marker
+    for m in _PAT_PRIVKEY.finditer(line):
+        v = m.group(0)
+        _claim(v, 5, _mk(path, lineno, "PROOFCTL-SECRET-005",
+                         "Private key", Severity.ERROR,
+                         "Private key block detected"))
+
+    # SECRET-006: JWT (warning — often in tests/docs)
+    for m in _PAT_JWT.finditer(line):
+        v = m.group(0)
+        _claim(v, 6, _mk(path, lineno, "PROOFCTL-SECRET-006",
+                         "JWT token", Severity.WARNING,
+                         f"JWT token detected ({v[:12]}…) — confirm not a real credential"))
+
+    # SECRET-008: Google API key
+    for m in _PAT_GOOGLE.finditer(line):
+        v = m.group(0)
+        _claim(v, 8, _mk(path, lineno, "PROOFCTL-SECRET-008",
+                         "Google API key", Severity.ERROR,
+                         f"Hardcoded Google API key detected ({v[:8]}…)"))
+
+    # SECRET-009: Stripe key
+    for m in _PAT_STRIPE.finditer(line):
+        v = m.group(0)
+        _claim(v, 9, _mk(path, lineno, "PROOFCTL-SECRET-009",
+                         "Stripe API key", Severity.ERROR,
+                         f"Hardcoded Stripe key detected ({v[:8]}…)"))
+
+    # SECRET-010: Terraform/HCL admin password
+    for m in _PAT_TF_ADMIN.finditer(line):
+        v = m.group("val")
+        if _is_env_ref(v) or _is_placeholder(v):
+            continue
+        _claim(v, 10, _mk(path, lineno, "PROOFCTL-SECRET-010",
+                          "Hardcoded admin password", Severity.ERROR,
+                          f"Hardcoded admin password in '{m.group('key')}' assignment"))
+
+    # SECRET-007: Generic high-entropy assignment (lowest priority — fires last)
+    for m in _PAT_GENERIC.finditer(line):
+        v = m.group("val")
+        if _is_env_ref(v) or _is_placeholder(v):
+            continue
+        if v.islower() and len(v) <= 16 and re.fullmatch(r"[a-z0-9_\-]+", v):
+            # Looks like a slug/placeholder — skip
+            continue
+        if _shannon_entropy(v) < 3.5:
+            continue
+        _claim(v, 7, _mk(path, lineno, "PROOFCTL-SECRET-007",
+                         "Hardcoded credential assignment", Severity.ERROR,
+                         f"Likely hardcoded {m.group('key')} assignment"))
+
+    out.extend(f for _, f in by_value.values())
+    return out
+
+
+# ── SECRET-011: kind: Secret YAML ─────────────────────────────────────────────
+
+def _scan_k8s_secrets(path: Path, source: str) -> list[Finding]:
+    """Scan YAML for `kind: Secret` documents with plaintext data:/stringData:.
+
+    Fires once per Secret document on the line of `kind: Secret`.
+    Also fires on multi-doc YAML (--- separators) per document.
+    """
+    name_str = str(path).replace("\\", "/").lower()
+    if any(tok in name_str for tok in _K8S_SKIP_PATH_TOKENS):
+        return []
+    # Restrict to plausible YAML files (cheap guard — main.yaml, *.yml, *.yaml).
+    if not (path.name.endswith(".yaml") or path.name.endswith(".yml")):
+        return []
+
+    lines = source.splitlines()
+    out: list[Finding] = []
+
+    # Collect document slices (line ranges) using --- separators.
+    doc_starts: list[int] = [0]
+    for idx, line in enumerate(lines):
+        if idx > 0 and _PAT_YAML_DOC_SEP.match(line):
+            doc_starts.append(idx + 1)
+    doc_starts.append(len(lines))
+
+    for di in range(len(doc_starts) - 1):
+        start, end = doc_starts[di], doc_starts[di + 1]
+        kind_line = -1
+        secret_name = "<unnamed>"
+        # Find kind: Secret in this doc
+        for i in range(start, end):
+            if _PAT_K8S_KIND_SECRET.match(lines[i]):
+                kind_line = i
+                break
+        if kind_line < 0:
+            continue
+
+        # Find metadata.name (best-effort; first 'name:' under metadata or any 'name:' line).
+        for i in range(start, end):
+            mn = _PAT_K8S_NAME.match(lines[i])
+            if mn:
+                secret_name = mn.group(1)
+                break
+
+        # Walk for data:/stringData: blocks and check for any non-empty entries.
+        i = start
+        leaked = False
+        while i < end:
+            line = lines[i]
+            mkey = _PAT_K8S_DATA_KEY.match(line)
+            if not mkey:
+                i += 1
+                continue
+            base_indent = len(mkey.group(1))
+            j = i + 1
+            while j < end:
+                child = lines[j]
+                if not child.strip() or child.lstrip().startswith("#"):
+                    j += 1
+                    continue
+                child_indent = len(child) - len(child.lstrip())
+                if child_indent <= base_indent:
+                    break
+                me = _PAT_K8S_DATA_ENTRY.match(child)
+                if me:
+                    val = me.group(3).strip()
+                    # Strip simple quoting
+                    if (val.startswith('"') and val.endswith('"')) or (
+                        val.startswith("'") and val.endswith("'")
+                    ):
+                        val = val[1:-1]
+                    # Skip empty / null / valueFrom-style references / pipe-block markers
+                    if not val or val in ("''", '""', "null", "~"):
+                        j += 1
+                        continue
+                    if val in ("|", ">", "|-", ">-", "|+", ">+"):
+                        # Block scalar follows — treat as a real value if any next
+                        # non-empty indented line exists.
+                        k = j + 1
+                        has_content = False
+                        while k < end:
+                            nxt = lines[k]
+                            if not nxt.strip():
+                                k += 1
+                                continue
+                            nxt_indent = len(nxt) - len(nxt.lstrip())
+                            if nxt_indent <= child_indent:
+                                break
+                            has_content = True
+                            break
+                        if has_content:
+                            leaked = True
+                            break
+                        j += 1
+                        continue
+                    if val.startswith("valueFrom") or val.startswith("&") or val.startswith("*"):
+                        j += 1
+                        continue
+                    # Otherwise: a real plaintext value (b64 or stringData).
+                    leaked = True
+                    break
+                j += 1
+            if leaked:
+                break
+            i = j
+
+        if leaked:
+            out.append(Finding(
+                file=str(path),
+                line=kind_line + 1,
+                col=0,
+                rule_id="PROOFCTL-SECRET-011",
+                rule_name="Plaintext credentials in Kubernetes Secret manifest",
+                severity=Severity.ERROR,
+                message=(
+                    f"Kubernetes Secret '{secret_name}' has plaintext credentials "
+                    "in data:/stringData:"
+                ),
+                hint=(
+                    "Use external secret store: ExternalSecrets, Sealed Secrets, "
+                    "or valueFrom.secretKeyRef from a properly-managed Secret "
+                    "outside the manifest."
+                ),
+                authority="OWASP ASVS V2 / CWE-798 – Use of Hard-coded Credentials",
+                docs_url="https://cwe.mitre.org/data/definitions/798.html",
+            ))
+    return out
+
+
+# ── SECRET-012: secret-shaped env assignment inside heredoc ───────────────────
+
+def _scan_heredoc_secrets(
+    path: Path, source: str, already_fired_lines: set[int]
+) -> list[Finding]:
+    """Detect FOO_KEY/_TOKEN/_SECRET/_PASSWORD=value inside <<EOF heredocs."""
+    out: list[Finding] = []
+    heredoc_end: str | None = None
+    for i, line in enumerate(source.splitlines(), start=1):
+        if heredoc_end is None:
+            hm = _PAT_HEREDOC_OPEN.search(line)
+            if hm:
+                heredoc_end = hm.group(1)
+            continue
+        # Inside heredoc
+        stripped = line.strip()
+        if stripped == heredoc_end or stripped.rstrip("-~") == heredoc_end:
+            heredoc_end = None
+            continue
+        if i in already_fired_lines:
+            continue
+        m = _PAT_HEREDOC_SECRET.search(line)
+        if not m:
+            continue
+        key = m.group(1)
+        val = m.group(2)
+        if (val.startswith('"') and val.endswith('"')) or (
+            val.startswith("'") and val.endswith("'")
+        ):
+            val = val[1:-1]
+        if len(val) < 8:
+            continue
+        if _is_env_ref(val) or val.startswith("$"):
+            continue
+        if _is_placeholder(val):
+            continue
+        out.append(Finding(
+            file=str(path),
+            line=i,
+            col=0,
+            rule_id="PROOFCTL-SECRET-012",
+            rule_name="Secret in heredoc body",
+            severity=Severity.ERROR,
+            message=f"Hardcoded secret-shaped assignment '{key}=...' inside heredoc",
+            hint=(
+                "Heredocs (e.g. user_data) are committed to state and version "
+                "control. Inject via EC2 Instance Metadata, SSM Parameter Store, "
+                "Secrets Manager, or a templated runtime variable."
+            ),
+            authority="OWASP ASVS V2 / CWE-798 – Use of Hard-coded Credentials",
+            docs_url="https://cwe.mitre.org/data/definitions/798.html",
+        ))
+    return out
+
+
+# ── main checker ──────────────────────────────────────────────────────────────
+
+class SecretChecker(FileChecker):
+    """Scans any text file for hardcoded secrets."""
+
+    extensions: tuple[str, ...] = ()
+
+    def check(self, path: Path, source: str, tree=None) -> list[Finding]:  # type: ignore[override]
+        if _should_skip(path):
+            return []
+        # Binary sniff using the source's first 1024 chars encoded back.
+        head_bytes = source[:1024].encode("utf-8", errors="replace")
+        if _is_binary(head_bytes):
+            return []
+
+        findings: list[Finding] = []
+        for i, line in enumerate(source.splitlines(), start=1):
+            if not line:
+                continue
+            findings.extend(_scan_line(path, i, line))
+
+        # SECRET-011: K8s Secret YAML
+        findings.extend(_scan_k8s_secrets(path, source))
+
+        # SECRET-012: heredoc secrets (skip lines where SECRET-001..010 already fired)
+        already = {f.line for f in findings if f.rule_id != "PROOFCTL-SECRET-011"}
+        findings.extend(_scan_heredoc_secrets(path, source, already))
+
+        return findings