Build Skill-Augmented AI Agents with SkillNet for Search, Evaluation, Graph Analysis, and Task Planning

banner("5) Evaluate skills on 5 quality dimensions (quality gate)")
DIMS = ["safety", "completeness", "executability", "maintainability", "cost_awareness"]
LEVEL_SCORE = {"Excellent": 4, "Good": 3, "Fair": 2, "Poor": 1, "Bad": 0}
def evaluate(target):
   if USE_SDK and API_KEY:
       try:
           return client.evaluate(target=target)
       except Exception as e:
           print(f"  evaluate failed for {target}: {e!r}")
   return None
def mock_eval(name):
   import hashlib
   h = int(hashlib.md5(name.encode()).hexdigest(), 16)
   levels = ["Excellent", "Good", "Fair", "Poor"]
   return {d: {"level": levels[(h >> (i * 3)) % 4], "reason": "offline mock score"}
           for i, d in enumerate(DIMS)}
def gate_score(report):
   tot = sum(LEVEL_SCORE.get(report.get(d, {}).get("level", "Fair"), 2) for d in DIMS)
   return tot / (len(DIMS) * 4)
GATE_THRESHOLD = 0.55
targets = [s["skill_url"] for s in (kw_hits + vec_hits) if s["skill_url"]][:3] \
         or [i["name"] for i in inspected] or ["pdf-extractor", "chart-reader", "web-scraper"]
passed, scored = [], []
for t in targets:
   rep = evaluate(t)
   via = "LLM"
   if rep is None:
       rep, via = mock_eval(str(t)), "mock"
   score = gate_score(rep)
   scored.append((t, score, via))
   flags = " ".join(f"{d[:4]}={rep.get(d,{}).get('level','?')[:4]}" for d in DIMS)
   status = "PASS ✅" if score >= GATE_THRESHOLD else "FAIL ❌"
   print(f"  [{via:4}] {status} score={score:.2f}  {textwrap.shorten(str(t),46,placeholder="...")}")
   print(f"          {flags}")
   if score >= GATE_THRESHOLD:
       passed.append(t)
print(f"\n{len(passed)}/{len(targets)} skills passed the quality gate (threshold={GATE_THRESHOLD}).")
banner("6) Analyze relationships and draw the Skill Graph")
def analyze(skills_dir):
   if USE_SDK and API_KEY:
       try:
           return client.analyze(skills_dir=str(skills_dir))
       except Exception as e:
           print(f"  analyze failed: {e!r}")
   return None
rels = analyze(SKILLS_DIR)
if not rels:
   names = [i["name"] for i in inspected] or ["PDF_Parser", "Text_Summarizer",
                                              "Chart_Reader", "Web_Scraper"]
   while len(names) < 4:
       names.append(f"Skill_{len(names)}")
   rels = [
       {"source": names[0], "type": "compose_with", "target": names[1]},
       {"source": names[2], "type": "similar_to",   "target": names[0]},
       {"source": names[3], "type": "depend_on",    "target": names[1]},
       {"source": names[1], "type": "belong_to",    "target": names[2]},
   ]
   print("  (using offline mock relationships — set API_KEY for real analysis)")
for r in rels:
   print(f"  {r['source']} --[{r['type']}]--> {r['target']}")
try:
   import networkx as nx
   import matplotlib.pyplot as plt
   G = nx.DiGraph()
   COLORS = {"similar_to": "#4C9BE8", "belong_to": "#E8A14C",
             "compose_with": "#6BBF59", "depend_on": "#D45D79"}
   for r in rels:
       G.add_edge(r["source"], r["target"], type=r["type"])
   pos = nx.spring_layout(G, seed=42, k=1.2)
   plt.figure(figsize=(9, 6))
   nx.draw_networkx_nodes(G, pos, node_size=2200, node_color="#EDEDED", edgecolors="#444")
   nx.draw_networkx_labels(G, pos, font_size=9)
   for et, col in COLORS.items():
       edges = [(u, v) for u, v, d in G.edges(data=True) if d["type"] == et]
       if edges:
           nx.draw_networkx_edges(G, pos, edgelist=edges, edge_color=col,
                                  width=2, arrows=True, arrowsize=18,
                                  connectionstyle="arc3,rad=0.08")
   plt.legend(handles=[plt.Line2D([0], [0], color=c, lw=2, label=t)
                       for t, c in COLORS.items()], loc="best", fontsize=8)
   plt.title("SkillNet — Skill Relationship Graph")
   plt.axis("off"); plt.tight_layout()
   plt.savefig(WORKDIR / "skill_graph.png", dpi=130)
   plt.show()
   print(f"  graph saved -> {WORKDIR/'skill_graph.png'}")
except Exception as e:
   print(f"  graph drawing skipped: {e!r}")