dsh-plugin-capabilities 0.1.4 → 0.1.6

package/skills/skill-creator/scripts/run_loop.py

@@ -0,0 +1,328 @@
+#!/usr/bin/env python3
+"""Run the eval + improve loop until all pass or max iterations reached.
+
+Combines run_eval.py and improve_description.py in a loop, tracking history
+and returning the best description found. Supports train/test split to prevent
+overfitting.
+"""
+
+import argparse
+import json
+import random
+import sys
+import tempfile
+import time
+import webbrowser
+from pathlib import Path
+
+from scripts.generate_report import generate_html
+from scripts.improve_description import improve_description
+from scripts.run_eval import find_project_root, run_eval
+from scripts.utils import parse_skill_md
+
+
+def split_eval_set(eval_set: list[dict], holdout: float, seed: int = 42) -> tuple[list[dict], list[dict]]:
+    """Split eval set into train and test sets, stratified by should_trigger."""
+    random.seed(seed)
+
+    # Separate by should_trigger
+    trigger = [e for e in eval_set if e["should_trigger"]]
+    no_trigger = [e for e in eval_set if not e["should_trigger"]]
+
+    # Shuffle each group
+    random.shuffle(trigger)
+    random.shuffle(no_trigger)
+
+    # Calculate split points
+    n_trigger_test = max(1, int(len(trigger) * holdout))
+    n_no_trigger_test = max(1, int(len(no_trigger) * holdout))
+
+    # Split
+    test_set = trigger[:n_trigger_test] + no_trigger[:n_no_trigger_test]
+    train_set = trigger[n_trigger_test:] + no_trigger[n_no_trigger_test:]
+
+    return train_set, test_set
+
+
+def run_loop(
+    eval_set: list[dict],
+    skill_path: Path,
+    description_override: str | None,
+    num_workers: int,
+    timeout: int,
+    max_iterations: int,
+    runs_per_query: int,
+    trigger_threshold: float,
+    holdout: float,
+    model: str,
+    verbose: bool,
+    live_report_path: Path | None = None,
+    log_dir: Path | None = None,
+) -> dict:
+    """Run the eval + improvement loop."""
+    project_root = find_project_root()
+    name, original_description, content = parse_skill_md(skill_path)
+    current_description = description_override or original_description
+
+    # Split into train/test if holdout > 0
+    if holdout > 0:
+        train_set, test_set = split_eval_set(eval_set, holdout)
+        if verbose:
+            print(f"Split: {len(train_set)} train, {len(test_set)} test (holdout={holdout})", file=sys.stderr)
+    else:
+        train_set = eval_set
+        test_set = []
+
+    history = []
+    exit_reason = "unknown"
+
+    for iteration in range(1, max_iterations + 1):
+        if verbose:
+            print(f"\n{'='*60}", file=sys.stderr)
+            print(f"Iteration {iteration}/{max_iterations}", file=sys.stderr)
+            print(f"Description: {current_description}", file=sys.stderr)
+            print(f"{'='*60}", file=sys.stderr)
+
+        # Evaluate train + test together in one batch for parallelism
+        all_queries = train_set + test_set
+        t0 = time.time()
+        all_results = run_eval(
+            eval_set=all_queries,
+            skill_name=name,
+            description=current_description,
+            num_workers=num_workers,
+            timeout=timeout,
+            project_root=project_root,
+            runs_per_query=runs_per_query,
+            trigger_threshold=trigger_threshold,
+            model=model,
+        )
+        eval_elapsed = time.time() - t0
+
+        # Split results back into train/test by matching queries
+        train_queries_set = {q["query"] for q in train_set}
+        train_result_list = [r for r in all_results["results"] if r["query"] in train_queries_set]
+        test_result_list = [r for r in all_results["results"] if r["query"] not in train_queries_set]
+
+        train_passed = sum(1 for r in train_result_list if r["pass"])
+        train_total = len(train_result_list)
+        train_summary = {"passed": train_passed, "failed": train_total - train_passed, "total": train_total}
+        train_results = {"results": train_result_list, "summary": train_summary}
+
+        if test_set:
+            test_passed = sum(1 for r in test_result_list if r["pass"])
+            test_total = len(test_result_list)
+            test_summary = {"passed": test_passed, "failed": test_total - test_passed, "total": test_total}
+            test_results = {"results": test_result_list, "summary": test_summary}
+        else:
+            test_results = None
+            test_summary = None
+
+        history.append({
+            "iteration": iteration,
+            "description": current_description,
+            "train_passed": train_summary["passed"],
+            "train_failed": train_summary["failed"],
+            "train_total": train_summary["total"],
+            "train_results": train_results["results"],
+            "test_passed": test_summary["passed"] if test_summary else None,
+            "test_failed": test_summary["failed"] if test_summary else None,
+            "test_total": test_summary["total"] if test_summary else None,
+            "test_results": test_results["results"] if test_results else None,
+            # For backward compat with report generator
+            "passed": train_summary["passed"],
+            "failed": train_summary["failed"],
+            "total": train_summary["total"],
+            "results": train_results["results"],
+        })
+
+        # Write live report if path provided
+        if live_report_path:
+            partial_output = {
+                "original_description": original_description,
+                "best_description": current_description,
+                "best_score": "in progress",
+                "iterations_run": len(history),
+                "holdout": holdout,
+                "train_size": len(train_set),
+                "test_size": len(test_set),
+                "history": history,
+            }
+            live_report_path.write_text(generate_html(partial_output, auto_refresh=True, skill_name=name))
+
+        if verbose:
+            def print_eval_stats(label, results, elapsed):
+                pos = [r for r in results if r["should_trigger"]]
+                neg = [r for r in results if not r["should_trigger"]]
+                tp = sum(r["triggers"] for r in pos)
+                pos_runs = sum(r["runs"] for r in pos)
+                fn = pos_runs - tp
+                fp = sum(r["triggers"] for r in neg)
+                neg_runs = sum(r["runs"] for r in neg)
+                tn = neg_runs - fp
+                total = tp + tn + fp + fn
+                precision = tp / (tp + fp) if (tp + fp) > 0 else 1.0
+                recall = tp / (tp + fn) if (tp + fn) > 0 else 1.0
+                accuracy = (tp + tn) / total if total > 0 else 0.0
+                print(f"{label}: {tp+tn}/{total} correct, precision={precision:.0%} recall={recall:.0%} accuracy={accuracy:.0%} ({elapsed:.1f}s)", file=sys.stderr)
+                for r in results:
+                    status = "PASS" if r["pass"] else "FAIL"
+                    rate_str = f"{r['triggers']}/{r['runs']}"
+                    print(f"  [{status}] rate={rate_str} expected={r['should_trigger']}: {r['query'][:60]}", file=sys.stderr)
+
+            print_eval_stats("Train", train_results["results"], eval_elapsed)
+            if test_summary:
+                print_eval_stats("Test ", test_results["results"], 0)
+
+        if train_summary["failed"] == 0:
+            exit_reason = f"all_passed (iteration {iteration})"
+            if verbose:
+                print(f"\nAll train queries passed on iteration {iteration}!", file=sys.stderr)
+            break
+
+        if iteration == max_iterations:
+            exit_reason = f"max_iterations ({max_iterations})"
+            if verbose:
+                print(f"\nMax iterations reached ({max_iterations}).", file=sys.stderr)
+            break
+
+        # Improve the description based on train results
+        if verbose:
+            print(f"\nImproving description...", file=sys.stderr)
+
+        t0 = time.time()
+        # Strip test scores from history so improvement model can't see them
+        blinded_history = [
+            {k: v for k, v in h.items() if not k.startswith("test_")}
+            for h in history
+        ]
+        new_description = improve_description(
+            skill_name=name,
+            skill_content=content,
+            current_description=current_description,
+            eval_results=train_results,
+            history=blinded_history,
+            model=model,
+            log_dir=log_dir,
+            iteration=iteration,
+        )
+        improve_elapsed = time.time() - t0
+
+        if verbose:
+            print(f"Proposed ({improve_elapsed:.1f}s): {new_description}", file=sys.stderr)
+
+        current_description = new_description
+
+    # Find the best iteration by TEST score (or train if no test set)
+    if test_set:
+        best = max(history, key=lambda h: h["test_passed"] or 0)
+        best_score = f"{best['test_passed']}/{best['test_total']}"
+    else:
+        best = max(history, key=lambda h: h["train_passed"])
+        best_score = f"{best['train_passed']}/{best['train_total']}"
+
+    if verbose:
+        print(f"\nExit reason: {exit_reason}", file=sys.stderr)
+        print(f"Best score: {best_score} (iteration {best['iteration']})", file=sys.stderr)
+
+    return {
+        "exit_reason": exit_reason,
+        "original_description": original_description,
+        "best_description": best["description"],
+        "best_score": best_score,
+        "best_train_score": f"{best['train_passed']}/{best['train_total']}",
+        "best_test_score": f"{best['test_passed']}/{best['test_total']}" if test_set else None,
+        "final_description": current_description,
+        "iterations_run": len(history),
+        "holdout": holdout,
+        "train_size": len(train_set),
+        "test_size": len(test_set),
+        "history": history,
+    }
+
+
+def main():
+    parser = argparse.ArgumentParser(description="Run eval + improve loop")
+    parser.add_argument("--eval-set", required=True, help="Path to eval set JSON file")
+    parser.add_argument("--skill-path", required=True, help="Path to skill directory")
+    parser.add_argument("--description", default=None, help="Override starting description")
+    parser.add_argument("--num-workers", type=int, default=10, help="Number of parallel workers")
+    parser.add_argument("--timeout", type=int, default=30, help="Timeout per query in seconds")
+    parser.add_argument("--max-iterations", type=int, default=5, help="Max improvement iterations")
+    parser.add_argument("--runs-per-query", type=int, default=3, help="Number of runs per query")
+    parser.add_argument("--trigger-threshold", type=float, default=0.5, help="Trigger rate threshold")
+    parser.add_argument("--holdout", type=float, default=0.4, help="Fraction of eval set to hold out for testing (0 to disable)")
+    parser.add_argument("--model", required=True, help="Model for improvement")
+    parser.add_argument("--verbose", action="store_true", help="Print progress to stderr")
+    parser.add_argument("--report", default="auto", help="Generate HTML report at this path (default: 'auto' for temp file, 'none' to disable)")
+    parser.add_argument("--results-dir", default=None, help="Save all outputs (results.json, report.html, log.txt) to a timestamped subdirectory here")
+    args = parser.parse_args()
+
+    eval_set = json.loads(Path(args.eval_set).read_text())
+    skill_path = Path(args.skill_path)
+
+    if not (skill_path / "SKILL.md").exists():
+        print(f"Error: No SKILL.md found at {skill_path}", file=sys.stderr)
+        sys.exit(1)
+
+    name, _, _ = parse_skill_md(skill_path)
+
+    # Set up live report path
+    if args.report != "none":
+        if args.report == "auto":
+            timestamp = time.strftime("%Y%m%d_%H%M%S")
+            live_report_path = Path(tempfile.gettempdir()) / f"skill_description_report_{skill_path.name}_{timestamp}.html"
+        else:
+            live_report_path = Path(args.report)
+        # Open the report immediately so the user can watch
+        live_report_path.write_text("<html><body><h1>Starting optimization loop...</h1><meta http-equiv='refresh' content='5'></body></html>")
+        webbrowser.open(str(live_report_path))
+    else:
+        live_report_path = None
+
+    # Determine output directory (create before run_loop so logs can be written)
+    if args.results_dir:
+        timestamp = time.strftime("%Y-%m-%d_%H%M%S")
+        results_dir = Path(args.results_dir) / timestamp
+        results_dir.mkdir(parents=True, exist_ok=True)
+    else:
+        results_dir = None
+
+    log_dir = results_dir / "logs" if results_dir else None
+
+    output = run_loop(
+        eval_set=eval_set,
+        skill_path=skill_path,
+        description_override=args.description,
+        num_workers=args.num_workers,
+        timeout=args.timeout,
+        max_iterations=args.max_iterations,
+        runs_per_query=args.runs_per_query,
+        trigger_threshold=args.trigger_threshold,
+        holdout=args.holdout,
+        model=args.model,
+        verbose=args.verbose,
+        live_report_path=live_report_path,
+        log_dir=log_dir,
+    )
+
+    # Save JSON output
+    json_output = json.dumps(output, indent=2)
+    print(json_output)
+    if results_dir:
+        (results_dir / "results.json").write_text(json_output)
+
+    # Write final HTML report (without auto-refresh)
+    if live_report_path:
+        live_report_path.write_text(generate_html(output, auto_refresh=False, skill_name=name))
+        print(f"\nReport: {live_report_path}", file=sys.stderr)
+
+    if results_dir and live_report_path:
+        (results_dir / "report.html").write_text(generate_html(output, auto_refresh=False, skill_name=name))
+
+    if results_dir:
+        print(f"Results saved to: {results_dir}", file=sys.stderr)
+
+
+if __name__ == "__main__":
+    main()

package/skills/skill-creator/scripts/utils.py

@@ -0,0 +1,47 @@
+"""Shared utilities for skill-creator scripts."""
+
+from pathlib import Path
+
+
+
+def parse_skill_md(skill_path: Path) -> tuple[str, str, str]:
+    """Parse a SKILL.md file, returning (name, description, full_content)."""
+    content = (skill_path / "SKILL.md").read_text()
+    lines = content.split("\n")
+
+    if lines[0].strip() != "---":
+        raise ValueError("SKILL.md missing frontmatter (no opening ---)")
+
+    end_idx = None
+    for i, line in enumerate(lines[1:], start=1):
+        if line.strip() == "---":
+            end_idx = i
+            break
+
+    if end_idx is None:
+        raise ValueError("SKILL.md missing frontmatter (no closing ---)")
+
+    name = ""
+    description = ""
+    frontmatter_lines = lines[1:end_idx]
+    i = 0
+    while i < len(frontmatter_lines):
+        line = frontmatter_lines[i]
+        if line.startswith("name:"):
+            name = line[len("name:"):].strip().strip('"').strip("'")
+        elif line.startswith("description:"):
+            value = line[len("description:"):].strip()
+            # Handle YAML multiline indicators (>, |, >-, |-)
+            if value in (">", "|", ">-", "|-"):
+                continuation_lines: list[str] = []
+                i += 1
+                while i < len(frontmatter_lines) and (frontmatter_lines[i].startswith("  ") or frontmatter_lines[i].startswith("\t")):
+                    continuation_lines.append(frontmatter_lines[i].strip())
+                    i += 1
+                description = " ".join(continuation_lines)
+                continue
+            else:
+                description = value.strip('"').strip("'")
+        i += 1
+
+    return name, description, content

package/src/client/CapabilitiesSection.tsx

@@ -0,0 +1,102 @@
+/** Settings top-level “技能与 MCP” section: one nav entry beside 通用设置 and
+ * 模型, with the Skills and MCP pages as internal tabs. The two pages are the
+ * same components the Plugins section used to host — promoted a level instead
+ * of squeezed under 插件. Pure composition; data still arrives through the
+ * injected faces. */
+
+import { useEffect, useId, useRef, useState } from 'react'
+import type { ReactElement } from 'react'
+import { CSS } from './css.ts'
+import { McpTab } from './McpTab.tsx'
+import type { McpInjected } from './McpTab.tsx'
+import { SkillsTab } from './SkillsTab.tsx'
+import type { SkillsInjected } from './SkillsTab.tsx'
+import type { Translate } from './index.ts'
+
+export function CapabilitiesSection(props: {
+  t: Translate
+  skills: SkillsInjected
+  mcp: McpInjected
+}): ReactElement {
+  const { t, skills, mcp } = props
+  const tabsId = useId()
+  const tabRefs = useRef<Array<HTMLButtonElement | null>>([])
+  const rows = [
+    { id: 'skills', label: t('skillsTab') },
+    { id: 'mcp', label: t('mcpTab') },
+  ]
+  const [activeId, setActiveId] = useState('skills')
+  // The skills page mounts immediately; a tab mounts only when first
+  // selected, then stays mounted while hidden so editor drafts and outcome
+  // banners survive switching between the two views.
+  const [visitedIds, setVisitedIds] = useState<ReadonlySet<string>>(() => new Set(['skills']))
+
+  useEffect(() => {
+    setVisitedIds((previous) => {
+      if (previous.has(activeId)) return previous
+      return new Set([...previous, activeId])
+    })
+  }, [activeId])
+
+  return (
+    <div className="dpc-section">
+      <style>{CSS}</style>
+
+      <h2 className="dpc-heading">{t('sectionNav')}</h2>
+      <div className="dpc-tabs" role="tablist" aria-label={t('sectionNav')}>
+        {rows.map((row, index) => {
+          const selected = row.id === activeId
+          return (
+            <button
+              key={row.id}
+              ref={(element) => { tabRefs.current[index] = element }}
+              id={`${tabsId}-tab-${row.id}`}
+              type="button"
+              role="tab"
+              className="dpc-tab"
+              aria-selected={selected}
+              aria-controls={`${tabsId}-panel-${row.id}`}
+              data-active={selected ? 'true' : undefined}
+              tabIndex={selected ? 0 : -1}
+              onClick={() => { setActiveId(row.id) }}
+              onKeyDown={(event) => {
+                let nextIndex: number
+                switch (event.key) {
+                  case 'ArrowRight': nextIndex = (index + 1) % rows.length; break
+                  case 'ArrowLeft': nextIndex = (index - 1 + rows.length) % rows.length; break
+                  case 'Home': nextIndex = 0; break
+                  case 'End': nextIndex = rows.length - 1; break
+                  default: return
+                }
+                event.preventDefault()
+                setActiveId(rows[nextIndex]?.id ?? 'skills')
+                tabRefs.current[nextIndex]?.focus()
+              }}
+            >
+              {row.label}
+            </button>
+          )
+        })}
+      </div>
+      {rows
+        .filter(row => row.id === activeId || visitedIds.has(row.id))
+        .map((row) => {
+          const selected = row.id === activeId
+          return (
+            <div
+              key={row.id}
+              id={`${tabsId}-panel-${row.id}`}
+              className="dpc-tabPanel"
+              role="tabpanel"
+              aria-labelledby={`${tabsId}-tab-${row.id}`}
+              hidden={!selected}
+            >
+              {row.id === 'skills'
+                ? <SkillsTab t={t} injected={skills} />
+                : <McpTab t={t} injected={mcp} />}
+            </div>
+          )
+        })}
+    </div>
+  )
+}

package/src/client/McpTab.tsx

@@ -320,6 +320,8 @@ export function McpTab(props: { t: Translate; injected: McpInjected }): ReactEle
         open={editor !== null}
         onClose={() => setEditor(null)}
         title={editor !== null && editor.id !== '' ? t('editServer') : t('addServer')}
+        className="dpc-modalForm"
+        contentClassName="dpc-modalScroll"
       >
         {editor !== null && (
           <div className="dpc-form">

package/src/client/SkillsTab.tsx

@@ -213,6 +213,8 @@ export function SkillsTab(props: { t: Translate; injected: SkillsInjected }): Re
         open={editor !== null}
         onClose={() => setEditor(null)}
         title={editor === null ? '' : editor.mode === 'create' ? t('newSkill') : editor.mode === 'edit' ? t('editSkill') : t('viewSkill')}
+        className="dpc-modalForm"
+        contentClassName="dpc-modalScroll"
       >
         {editor !== null && (
           <div className="dpc-form">

package/src/client/css.ts

@@ -4,6 +4,15 @@
 
 export const CSS = `
 .dpc-section{display:flex;flex-direction:column;gap:14px;width:100%;max-width:760px;color:var(--dsw-alias-label-primary)}
+.dpc-heading{margin:0;font-size:18px;line-height:26px;font-weight:600}
+/* Top-level section's internal tabs — same underline-tab look as the host's
+   Plugins section so the promoted placement still reads as one family. */
+.dpc-tabs{display:flex;align-items:flex-end;gap:22px;border-bottom:1px solid var(--dsw-alias-border-l2);margin-top:2px}
+.dpc-tab{position:relative;border:0;padding:7px 1px 9px;background:transparent;color:var(--dsw-alias-label-tertiary);font:inherit;font-size:13px;line-height:20px;cursor:pointer}
+.dpc-tab:hover,.dpc-tab[data-active='true']{color:var(--dsw-alias-label-primary)}
+.dpc-tab[data-active='true']::after,.dpc-tab:focus-visible::after{position:absolute;right:0;bottom:-1px;left:0;height:2px;border-radius:2px 2px 0 0;background:var(--dsw-alias-label-primary);content:''}
+.dpc-tab:focus-visible{outline:2px solid var(--dsw-alias-state-business-primary);outline-offset:2px;border-radius:2px;color:var(--dsw-alias-label-primary)}
+.dpc-tabPanel{min-width:0;padding-top:2px}
 .dpc-head{display:flex;align-items:center;gap:8px}
 .dpc-head h3{margin:0;font-size:13px;line-height:20px;font-weight:600}
 .dpc-head>svg{flex:none;color:var(--dsw-alias-label-tertiary)}
@@ -37,8 +46,8 @@ export const CSS = `
 .dpc-label{display:flex;flex-direction:column;gap:4px;font-size:12px;line-height:18px;color:var(--dsw-alias-label-secondary)}
 .dpc-label>span:first-child{color:var(--dsw-alias-label-tertiary)}
 .dpc-input,.dpc-textarea,.dpc-select{width:100%;box-sizing:border-box;border:1px solid var(--dsw-alias-border-l2);border-radius:8px;padding:7px 10px;outline:none;background:var(--dsw-alias-bg-layer-1);color:var(--dsw-alias-label-primary);font:inherit;font-size:13px}
-.dpc-textarea{min-height:120px;resize:vertical;font-family:var(--ds-font-family-code);line-height:1.5}
-.dpc-textarea[data-short='true']{min-height:64px}
+.dpc-textarea{min-height:240px;resize:vertical;font-family:var(--ds-font-family-code);line-height:1.5}
+.dpc-textarea[data-short='true']{min-height:96px}
 .dpc-input:focus-visible,.dpc-textarea:focus-visible,.dpc-select:focus-visible{border-color:var(--dsw-alias-state-business-primary);box-shadow:0 0 0 2px color-mix(in srgb,var(--dsw-alias-state-business-primary) 18%,transparent)}
 .dpc-checks{display:flex;gap:16px;font-size:13px;line-height:20px}
 .dpc-checks label{display:inline-flex;align-items:center;gap:6px;cursor:pointer}
@@ -47,6 +56,11 @@ export const CSS = `
    its own scroll so title, intro, and footer stay pinned. The doubled class
    beats the host dialog's module CSS regardless of injection order. */
 .dpc-modalWide.dpc-modalWide{width:min(680px,100%)}
+/* Editor dialogs (new skill / server): 640px wide so markdown bodies and
+   command/arg/env lines stop wrapping mid-token; the content column scrolls
+   on short viewports instead of clipping past the dialog edge. */
+.dpc-modalForm.dpc-modalForm{width:min(640px,100%)}
+.dpc-modalScroll.dpc-modalScroll{max-height:calc(100vh - 160px);overflow-y:auto}
 .dpc-importScroll{display:flex;flex-direction:column;gap:14px;max-height:min(400px,52vh);overflow-y:auto;padding:2px 4px 2px 2px}
 .dpc-importGroup{display:flex;flex-direction:column;gap:8px}
 .dpc-importHead{display:flex;align-items:center;gap:8px;padding:0 2px}

package/src/client/index.ts

@@ -1,9 +1,11 @@
-/** dsh-plugin-capabilities client entry: contributes the “技能/Skills” and
- * “MCP” tabs into Settings → Plugins. Calls the host routes with fetch. */
+/** dsh-plugin-capabilities client entry: contributes the top-level
+ * “技能与 MCP” section into Settings (beside 通用设置/模型, with Skills and
+ * MCP as internal tabs). Calls the host routes with fetch. */
 
 import { createElement as h } from 'react'
-import { McpTab, type McpInjected, type McpRow } from './McpTab.tsx'
-import { SkillsTab, type SkillsInjected, type SkillRowView } from './SkillsTab.tsx'
+import { CapabilitiesSection } from './CapabilitiesSection.tsx'
+import type { McpInjected, McpRow } from './McpTab.tsx'
+import type { SkillsInjected, SkillRowView } from './SkillsTab.tsx'
 import { zh, en } from './locales.ts'
 
 /** Locale dictionary namespace owned by this plugin. */
@@ -91,25 +93,16 @@ export function apply(ctx: CapabilitiesClientContext): void {
     desktop: window.dshDesktop !== undefined,
   }
 
-  ctx.slots.inject('settings.plugins.tab', () => {
+  // One top-level nav entry (between 模型 and 插件), not a tab under the
+  // Plugins section — the section component owns its internal Skills/MCP
+  // tabs directly, so nothing registers into settings.plugins.tab anymore.
+  ctx.slots.inject('settings.section', () => {
     return ctx.slots.register({
-      name: 'settings.plugins.tab',
-      id: 'capabilities-skills',
-      order: 30,
-      label: () => t('skillsTab'),
+      name: 'settings.section',
+      id: 'capabilities',
+      order: 12,
+      label: () => t('sectionNav'),
       locale: NS,
-      inject: () => skillsInjected,
-    }, () => h(SkillsTab, { t, injected: skillsInjected }))
-  })
-
-  ctx.slots.inject('settings.plugins.tab', () => {
-    return ctx.slots.register({
-      name: 'settings.plugins.tab',
-      id: 'capabilities-mcp',
-      order: 40,
-      label: () => t('mcpTab'),
-      locale: NS,
-      inject: () => mcpInjected,
-    }, () => h(McpTab, { t, injected: mcpInjected }))
+    }, () => h(CapabilitiesSection, { t, skills: skillsInjected, mcp: mcpInjected }))
   })
 }

package/src/client/locales.ts

@@ -1,6 +1,7 @@
 /** zh/en dictionaries for the Settings capabilities tabs. */
 
 export const zh = {
+  sectionNav: '技能与 MCP',
   skillsTab: '技能',
   mcpTab: 'MCP',
   skillsTitle: '技能管理',
@@ -72,6 +73,7 @@ export const zh = {
 }
 
 export const en = {
+  sectionNav: 'Skills & MCP',
   skillsTab: 'Skills',
   mcpTab: 'MCP',
   skillsTitle: 'Skills',

package/src/index.ts

@@ -4,6 +4,9 @@
  * catalog (the web composition deliberately leaves the host row to presets). */
 
 import type { Context } from '@deepseek-ai/cordis'
+import { existsSync } from 'node:fs'
+import { dirname, join } from 'node:path'
+import { fileURLToPath } from 'node:url'
 import { agentSkillRoots } from './agents.ts'
 import { argvProfile, profileDir } from './profile.ts'
 import { mountCapabilitiesRoutes } from './routes.ts'
@@ -11,6 +14,16 @@ import type { CapabilitiesHost } from './types.ts'
 
 export const name = 'dsh-plugin-capabilities'
 
+/**
+ * The package's own vendored skills (`skills/` at the package root — resolves
+ * identically from src/ under vitest and from lib/ when installed). Scanned as
+ * a custom root, so every session sees them through the registry's global
+ * layer while the files stay zero-copy and travel with plugin installs.
+ */
+export function packagedSkillsDir(): string {
+  return join(dirname(fileURLToPath(import.meta.url)), '..', 'skills')
+}
+
 /** Optional cordis.yml configuration; profile defaults to the booted one. */
 export interface Config {
   /** Profile whose patch layer holds the MCP rows; defaults to argv or `web`. */
@@ -31,17 +44,20 @@ export function apply(ctx: Context, config?: Config): void {
     // The web bundle disables the host-plane `skill-filesystem` row on
     // purpose (presets own per-session discovery). The Settings manager
     // mounts its own host-plane provider as a CHILD of this plugin: it dies
-    // with us, registers into the registry's global layer, and preset layers
-    // keep their semantics (nearest layer still wins duplicate names). Other
-    // agents' skill roots (~/.claude/skills, ~/.codex/skills) join as custom
-    // dirs — zero-copy, live-synced both ways. A failed load only means an
-    // empty catalog — the routes keep serving.
+    // with us, registers into the registry's global layer (deployment-level
+    // providers are exactly what that layer is for — agents read the merged
+    // catalog), and preset layers keep their semantics (nearest layer still
+    // wins duplicate names). Custom roots, in scan order: the package's own
+    // vendored skills (read-only, update with the plugin), then other agents'
+    // skill roots (~/.claude/skills, ~/.codex/skills) — zero-copy, live-synced
+    // both ways. A failed load only means an empty catalog — the routes keep
+    // serving.
     void (async () => {
       try {
         const mod = (await import('@deepseek-ai/dsh-skill-filesystem')) as unknown as
           (FilesystemSkillPlugin & { default?: FilesystemSkillPlugin })
         const plugin = mod.default ?? mod
-        const roots = agentSkillRoots()
+        const roots = [packagedSkillsDir(), ...agentSkillRoots()].filter(dir => existsSync(dir))
         hostCtx.plugin(plugin, roots.length > 0 ? { customSkillDirs: roots } : {})
       } catch {
         // Unresolvable provider: skills list stays empty; MCP tab unaffected.