@event4u/agent-config 3.3.0 → 4.1.0

package/scripts/print_required_checks.py

@@ -0,0 +1,196 @@
+#!/usr/bin/env python3
+"""Print the expected required-check set for the current branch.
+
+Contract: `docs/contracts/branch-protection-policy.md`. Per-PR-shape
+required-check floor — feature PR vs release PR vs docs-only PR. The
+script resolves the PR shape locally so the maintainer can sanity-check
+before pushing, without round-tripping through the GitHub UI.
+
+Resolution order:
+
+  1. `--branch <name>` flag — explicit override.
+  2. Current git branch — `git rev-parse --abbrev-ref HEAD`.
+  3. Fail with exit 2 (usage error).
+
+PR-shape classification:
+
+  release   — branch matches `^release/\\d+\\.\\d+\\.\\d+$` AND
+              `check_release_pr_shape.py` reports SHAPE-CLEAN for the
+              local diff against `--base` (default `origin/main`).
+  docs-only — diff vs base is entirely inside `docs/**` or top-level
+              `.md` files (`README.md`, `CHANGELOG.md`,
+              `CONTRIBUTING.md`, `AGENTS.md`).
+  feature   — everything else (default).
+
+The script never invokes `gh` and never touches the network — it works
+offline against the local git index so pre-push previews stay fast.
+
+Exit codes:
+
+  0 — printed the expected required-check set.
+  1 — release-PR shape detector reported OUT-OF-SHAPE; falls back to
+      the feature-PR set, which is also printed, plus a warning.
+  2 — usage / environment error.
+"""
+
+from __future__ import annotations
+
+import argparse
+import re
+import subprocess
+import sys
+from pathlib import Path
+
+REPO_ROOT = Path(__file__).resolve().parent.parent
+sys.path.insert(0, str(REPO_ROOT / "scripts"))
+
+import check_release_pr_shape as shape  # noqa: E402
+
+RELEASE_BRANCH_RE = re.compile(r"^release/\d+\.\d+\.\d+$")
+
+DOCS_ONLY_ALLOWED_TOP = {
+    "README.md",
+    "CHANGELOG.md",
+    "CONTRIBUTING.md",
+    "AGENTS.md",
+    "LICENSE",
+    "llms.txt",
+}
+
+FEATURE_CHECKS = (
+    "Consistency",
+    "Smoke Contracts",
+    "Skill Lint",
+    "Tests / install-tests",
+    "Tests / install-aux-tests",
+    "Tests / python-tests",
+    "Tests / node-tests",
+    "Public Install Smoke / smoke",
+)
+
+RELEASE_CHECKS = (
+    "Consistency",
+    "Smoke Contracts",
+    "Migration Dry-Run",
+    "Release Validation / release-shape",
+    "Release Validation / changelog-entry",
+    "Release Validation / version-consistency",
+    "Release Guard (post-tag)",
+)
+
+DOCS_ONLY_CHECKS = (
+    "Consistency",
+    "Smoke Contracts",
+)
+
+
+def _git(*args: str) -> str:
+    return subprocess.run(
+        ["git", *args], check=True, capture_output=True, text=True
+    ).stdout.strip()
+
+
+def current_branch() -> str:
+    return _git("rev-parse", "--abbrev-ref", "HEAD")
+
+
+def diff_files(base: str) -> list[str]:
+    out = subprocess.run(
+        ["git", "diff", "--name-only", f"{base}...HEAD"],
+        capture_output=True,
+        text=True,
+    )
+    if out.returncode != 0:
+        # Base ref unknown locally — fall back to staged + tracked
+        # changes vs HEAD so the dry-run still tells the maintainer
+        # something useful instead of crashing.
+        out = subprocess.run(
+            ["git", "diff", "--name-only", "HEAD"],
+            check=True,
+            capture_output=True,
+            text=True,
+        )
+    return [line for line in out.stdout.splitlines() if line.strip()]
+
+
+def is_docs_only(files: list[str]) -> bool:
+    if not files:
+        return False
+    for f in files:
+        if f in DOCS_ONLY_ALLOWED_TOP:
+            continue
+        if f.startswith("docs/"):
+            continue
+        return False
+    return True
+
+
+def classify(branch: str, files: list[str]) -> tuple[str, int]:
+    """Return (shape, exit_code)."""
+    if RELEASE_BRANCH_RE.match(branch):
+        # Release-shape predicate — re-uses the fail-closed allowlist.
+        bad = [f for f in files if not shape._matches(f)]
+        if bad:
+            print(
+                "WARNING: branch matches release/X.Y.Z but diff contains "
+                f"{len(bad)} out-of-allowlist file(s):",
+                file=sys.stderr,
+            )
+            for f in bad:
+                print(f"  - {f}", file=sys.stderr)
+            print(
+                "Falling back to feature-PR required-check set.",
+                file=sys.stderr,
+            )
+            return "feature", 1
+        return "release", 0
+    if is_docs_only(files):
+        return "docs-only", 0
+    return "feature", 0
+
+
+def print_set(shape_label: str, files: list[str]) -> None:
+    table = {
+        "feature": FEATURE_CHECKS,
+        "release": RELEASE_CHECKS,
+        "docs-only": DOCS_ONLY_CHECKS,
+    }[shape_label]
+    print(f"PR shape: {shape_label}  ({len(files)} file(s) in diff)")
+    print(f"Required checks ({len(table)}):")
+    for name in table:
+        print(f"  - {name}")
+    print()
+    print(
+        "Contract: docs/contracts/branch-protection-policy.md "
+        "(per-PR-shape matrix)"
+    )
+
+
+def main(argv: list[str] | None = None) -> int:
+    parser = argparse.ArgumentParser(
+        description=(
+            "Print the expected required-check set for the current branch."
+        ),
+    )
+    parser.add_argument(
+        "--branch",
+        help="Branch name override (default: current git branch).",
+    )
+    parser.add_argument(
+        "--base",
+        default="origin/main",
+        help="Base ref for the diff (default: origin/main).",
+    )
+    args = parser.parse_args(argv)
+
+    branch = args.branch or current_branch()
+    files = diff_files(args.base)
+    shape_label, exit_code = classify(branch, files)
+    print(f"Branch: {branch}")
+    print(f"Base:   {args.base}")
+    print_set(shape_label, files)
+    return exit_code
+
+
+if __name__ == "__main__":
+    sys.exit(main())

package/scripts/probe_skill_registration.py

@@ -0,0 +1,413 @@
+#!/usr/bin/env python3
+"""Tool-agnostic skill-registration probe.
+
+Roadmap: road-to-clean-skill-distribution-channels.md § Phase C.
+Contract: docs/contracts/skill-distribution-channels.md.
+
+Surfaces every skill registered for any of the six supported AI tools
+across user-global, project-local, and plugin-manifest sources. Flags
+``DUPLICATE`` (same skill name registered in ≥ 2 sources) and ``DRIFT``
+(same name, different description-hash or version). Both shapes are the
+class of bug that opened road-to-clean-skill-distribution-channels.md
+on 2026-05-25.
+
+CLI:
+
+    python3 scripts/probe_skill_registration.py
+    python3 scripts/probe_skill_registration.py --tool=claude --format=json
+    python3 scripts/probe_skill_registration.py --strict
+
+``--strict`` flips the exit code: 0 if no DUPLICATE / DRIFT findings,
+non-zero otherwise. Without ``--strict`` the script is informational
+(always exits 0).
+"""
+from __future__ import annotations
+
+import argparse
+import hashlib
+import json
+import os
+import sys
+from dataclasses import dataclass, field, asdict
+from pathlib import Path
+from typing import Iterable
+
+TOOL_IDS = ("claude", "augment", "cursor", "cline", "windsurf", "copilot")
+SCOPE_IDS = ("user", "project")
+
+
+@dataclass(frozen=True)
+class Registration:
+    """One row in the probe table — a single (skill_id, scope, source) tuple."""
+    skill_id: str
+    tool: str
+    scope: str
+    source_path: str
+    version: str
+    description_snippet: str
+    description_hash: str
+
+    def to_dict(self) -> dict[str, str]:
+        return asdict(self)
+
+
+@dataclass
+class ProbeResult:
+    registrations: list[Registration] = field(default_factory=list)
+    duplicates: dict[str, list[Registration]] = field(default_factory=dict)
+    drift: dict[str, list[Registration]] = field(default_factory=dict)
+
+    def to_dict(self) -> dict[str, object]:
+        return {
+            "registrations": [r.to_dict() for r in self.registrations],
+            "duplicates": {k: [r.to_dict() for r in v] for k, v in self.duplicates.items()},
+            "drift": {k: [r.to_dict() for r in v] for k, v in self.drift.items()},
+        }
+
+
+# ---------------------------------------------------------------------------
+# Frontmatter + version helpers
+# ---------------------------------------------------------------------------
+
+def _read_frontmatter(skill_md: Path) -> dict[str, str]:
+    """Minimal YAML frontmatter extractor — no PyYAML dependency."""
+    try:
+        text = skill_md.read_text(encoding="utf-8", errors="replace")
+    except OSError:
+        return {}
+    if not text.startswith("---\n") and not text.startswith("---\r\n"):
+        return {}
+    rest = text.split("---", 2)
+    if len(rest) < 3:
+        return {}
+    body = rest[1]
+    out: dict[str, str] = {}
+    for line in body.splitlines():
+        if ":" not in line or line.lstrip().startswith("#"):
+            continue
+        key, _, value = line.partition(":")
+        out[key.strip()] = value.strip().strip('"').strip("'")
+    return out
+
+
+def _hash_desc(desc: str) -> str:
+    return hashlib.sha256(desc.encode("utf-8", "replace")).hexdigest()[:12]
+
+
+def _snippet(desc: str, n: int = 80) -> str:
+    desc = desc.strip()
+    return desc if len(desc) <= n else desc[: n - 1] + "…"
+
+
+def _version_at(root: Path) -> str:
+    """Read a 'this install's version' value from package.json / plugin.json."""
+    for candidate in (root / "package.json", root / ".augment-plugin" / "plugin.json"):
+        if not candidate.is_file():
+            continue
+        try:
+            data = json.loads(candidate.read_text(encoding="utf-8"))
+            v = data.get("version")
+            if isinstance(v, str) and v:
+                return v
+        except (OSError, json.JSONDecodeError):
+            continue
+    return "unknown"
+
+
+# ---------------------------------------------------------------------------
+# Per-tool readers
+# ---------------------------------------------------------------------------
+
+def _iter_skill_md(skills_root: Path) -> Iterable[Path]:
+    if not skills_root.is_dir():
+        return
+    for entry in sorted(skills_root.iterdir()):
+        if not entry.is_dir():
+            continue
+        skill_md = entry / "SKILL.md"
+        if skill_md.is_file():
+            yield skill_md
+
+
+def _read_claude(scope: str, root: Path) -> Iterable[Registration]:
+    skills = root / ".claude" / "skills"
+    for skill_md in _iter_skill_md(skills):
+        fm = _read_frontmatter(skill_md)
+        name = fm.get("name") or skill_md.parent.name
+        desc = fm.get("description", "")
+        yield Registration(
+            skill_id=name,
+            tool="claude",
+            scope=scope,
+            source_path=str(skill_md),
+            version=_version_at(root),
+            description_snippet=_snippet(desc),
+            description_hash=_hash_desc(desc),
+        )
+    # Plugin manifest at the same scope. Each entry is a path string into
+    # the same .claude/skills/ tree; emit it as a separate "manifest" row so
+    # duplicate-detection catches manifest-vs-filesystem double-counting.
+    manifest = root / ".claude-plugin" / "marketplace.json"
+    if manifest.is_file():
+        try:
+            data = json.loads(manifest.read_text(encoding="utf-8"))
+            plugins = data.get("plugins") or []
+            entries: list[str] = []
+            for plug in plugins:
+                entries.extend(plug.get("skills") or [])
+            for entry in entries:
+                # The entry is a path string; resolve the name from the path.
+                tail = Path(entry).name
+                yield Registration(
+                    skill_id=tail,
+                    tool="claude",
+                    scope=f"{scope}-plugin",
+                    source_path=str(manifest),
+                    version=_version_at(root),
+                    description_snippet="(plugin manifest entry)",
+                    description_hash="manifest",
+                )
+        except (OSError, json.JSONDecodeError):
+            pass
+
+
+def _read_augment(scope: str, root: Path) -> Iterable[Registration]:
+    skills = root / ".augment" / "skills"
+    for skill_md in _iter_skill_md(skills):
+        fm = _read_frontmatter(skill_md)
+        name = fm.get("name") or skill_md.parent.name
+        desc = fm.get("description", "")
+        yield Registration(
+            skill_id=name,
+            tool="augment",
+            scope=scope,
+            source_path=str(skill_md),
+            version=_version_at(root),
+            description_snippet=_snippet(desc),
+            description_hash=_hash_desc(desc),
+        )
+
+
+def _read_cursor(scope: str, root: Path) -> Iterable[Registration]:
+    rules = root / ".cursor" / "rules"
+    if not rules.is_dir():
+        return
+    for rule in sorted(rules.glob("*.mdc")):
+        fm = _read_frontmatter(rule)
+        name = fm.get("name") or rule.stem
+        desc = fm.get("description", "")
+        yield Registration(
+            skill_id=name,
+            tool="cursor",
+            scope=scope,
+            source_path=str(rule),
+            version=_version_at(root),
+            description_snippet=_snippet(desc),
+            description_hash=_hash_desc(desc),
+        )
+
+
+def _read_cline(scope: str, root: Path) -> Iterable[Registration]:
+    rules = root / ".clinerules"
+    if not rules.is_dir():
+        return
+    for rule in sorted(rules.glob("*.md")):
+        fm = _read_frontmatter(rule)
+        name = fm.get("name") or rule.stem
+        desc = fm.get("description", "")
+        yield Registration(
+            skill_id=name,
+            tool="cline",
+            scope=scope,
+            source_path=str(rule),
+            version=_version_at(root),
+            description_snippet=_snippet(desc),
+            description_hash=_hash_desc(desc),
+        )
+
+
+def _read_windsurf(scope: str, root: Path) -> Iterable[Registration]:
+    rules = root / ".windsurf" / "rules"
+    if not rules.is_dir():
+        return
+    for rule in sorted(rules.glob("*.md")):
+        fm = _read_frontmatter(rule)
+        name = fm.get("name") or rule.stem
+        desc = fm.get("description", "")
+        yield Registration(
+            skill_id=name,
+            tool="windsurf",
+            scope=scope,
+            source_path=str(rule),
+            version=_version_at(root),
+            description_snippet=_snippet(desc),
+            description_hash=_hash_desc(desc),
+        )
+
+
+def _read_copilot(scope: str, root: Path) -> Iterable[Registration]:
+    # Copilot ships a single file; check both common locations.
+    for candidate in (
+        root / ".github" / "copilot-instructions.md",
+        root / "copilot-instructions.md",
+    ):
+        if not candidate.is_file():
+            continue
+        try:
+            text = candidate.read_text(encoding="utf-8", errors="replace")
+        except OSError:
+            continue
+        snippet = _snippet(text.split("\n", 1)[0] if text else "")
+        yield Registration(
+            skill_id="copilot-instructions",
+            tool="copilot",
+            scope=scope,
+            source_path=str(candidate),
+            version=_version_at(root),
+            description_snippet=snippet,
+            description_hash=_hash_desc(text),
+        )
+
+
+TOOL_READERS = {
+    "claude":   _read_claude,
+    "augment":  _read_augment,
+    "cursor":   _read_cursor,
+    "cline":    _read_cline,
+    "windsurf": _read_windsurf,
+    "copilot":  _read_copilot,
+}
+
+
+# ---------------------------------------------------------------------------
+# Probe
+# ---------------------------------------------------------------------------
+
+def run_probe(
+    *,
+    tool_filter: str = "all",
+    scope_filter: str = "all",
+    home: Path | None = None,
+    project: Path | None = None,
+) -> ProbeResult:
+    home = home or Path(os.environ.get("HOME", "/tmp"))
+    project = project or Path.cwd()
+
+    scopes: dict[str, Path] = {}
+    if scope_filter in ("all", "user"):
+        scopes["user"] = home
+    if scope_filter in ("all", "project"):
+        scopes["project"] = project
+
+    tools = TOOL_IDS if tool_filter == "all" else (tool_filter,)
+
+    result = ProbeResult()
+    for scope, root in scopes.items():
+        for tool in tools:
+            reader = TOOL_READERS.get(tool)
+            if reader is None:
+                continue
+            for reg in reader(scope, root):
+                result.registrations.append(reg)
+
+    # Group by (tool, skill_id). Anything with ≥ 2 entries is DUPLICATE.
+    # If their description hashes diverge, also flag DRIFT.
+    by_key: dict[tuple[str, str], list[Registration]] = {}
+    for reg in result.registrations:
+        by_key.setdefault((reg.tool, reg.skill_id), []).append(reg)
+
+    for (tool, skill_id), regs in by_key.items():
+        if len(regs) < 2:
+            continue
+        key = f"{tool}:{skill_id}"
+        result.duplicates[key] = regs
+        hashes = {r.description_hash for r in regs if r.description_hash != "manifest"}
+        versions = {r.version for r in regs if r.version != "unknown"}
+        if len(hashes) > 1 or len(versions) > 1:
+            result.drift[key] = regs
+
+    return result
+
+
+# ---------------------------------------------------------------------------
+# Rendering
+# ---------------------------------------------------------------------------
+
+def render_text(result: ProbeResult) -> str:
+    lines: list[str] = []
+    lines.append("Skill-registration probe")
+    lines.append("=" * 64)
+    lines.append("")
+    if not result.registrations:
+        lines.append("(no registrations found)")
+        return "\n".join(lines)
+
+    lines.append(f"{'TOOL':<10} {'SCOPE':<14} {'SKILL':<32} {'VER':<10} SOURCE")
+    lines.append("-" * 100)
+    for reg in result.registrations:
+        lines.append(
+            f"{reg.tool:<10} {reg.scope:<14} {reg.skill_id[:32]:<32} {reg.version:<10} {reg.source_path}"
+        )
+
+    if result.duplicates:
+        lines.append("")
+        lines.append("DUPLICATE — same skill registered in ≥ 2 sources")
+        lines.append("-" * 64)
+        for key, regs in result.duplicates.items():
+            lines.append(f"  {key}")
+            for r in regs:
+                lines.append(f"    - [{r.scope}/{r.version}] {r.source_path}")
+    if result.drift:
+        lines.append("")
+        lines.append("DRIFT — same skill registered with DIFFERENT description / version")
+        lines.append("-" * 64)
+        for key, regs in result.drift.items():
+            lines.append(f"  {key}")
+            for r in regs:
+                lines.append(f"    - [{r.scope}/{r.version}] hash={r.description_hash} desc={r.description_snippet!r}")
+                lines.append(f"      source: {r.source_path}")
+
+    return "\n".join(lines)
+
+
+def render_json(result: ProbeResult) -> str:
+    return json.dumps(result.to_dict(), indent=2)
+
+
+# ---------------------------------------------------------------------------
+# CLI
+# ---------------------------------------------------------------------------
+
+def main(argv: list[str] | None = None) -> int:
+    parser = argparse.ArgumentParser(
+        prog="probe_skill_registration",
+        description="Detect duplicate / drifting skill registrations across AI tools.",
+    )
+    parser.add_argument("--tool", choices=("all", *TOOL_IDS), default="all")
+    parser.add_argument("--scope", choices=("all", *SCOPE_IDS), default="all")
+    parser.add_argument("--format", choices=("text", "json"), default="text")
+    parser.add_argument(
+        "--strict", action="store_true",
+        help="Exit non-zero if any DUPLICATE / DRIFT finding is present.",
+    )
+    parser.add_argument("--home", type=Path, default=None, help="Override the user-global root (testing).")
+    parser.add_argument("--project", type=Path, default=None, help="Override the project root (testing).")
+    args = parser.parse_args(argv)
+
+    result = run_probe(
+        tool_filter=args.tool,
+        scope_filter=args.scope,
+        home=args.home,
+        project=args.project,
+    )
+
+    out = render_json(result) if args.format == "json" else render_text(result)
+    print(out)
+
+    if args.strict and (result.duplicates or result.drift):
+        return 2
+    return 0
+
+
+if __name__ == "__main__":
+    sys.exit(main())