@dsalazardev/ai-hub-cli 0.0.0-stage → 0.1.0

package/skills/shared/building-agents/SKILL.md

@@ -0,0 +1,386 @@
+---
+name: building-agents
+description: "Use when building or restructuring an LLM agent — provider adapter, tool calling, structured output, RAG, agent loop, eval gate, cost routing, tracing, MCP server — model-agnostic across OpenAI/Anthropic/Gemini/OSS so a model swap is a config change. NOT vector-store SQL alone (that is `postgresdb`) or service deployment (that is `deployment`)."
+tags: [agents, llm, mcp, rag, evals, ai]
+recommends: [secure-coding, deployment]
+origin: ai-hub
+---
+
+# Building production LLM agents (model-agnostic)
+
+A thin provider adapter, a disciplined agent loop, schema-validated tools, provider-neutral RAG, eval gates, OTel tracing, and optionally an MCP server — so swapping OpenAI ↔ Anthropic ↔ Gemini ↔ OSS is a config change, not a rewrite.
+
+## The one rule
+
+> Program against a **capability interface**, never a vendor SDK. Vendor specifics (model id, tool-schema shape, JSON mode, caching, token limits) live behind one adapter resolved from config. Model names and prices rot — if one appears in business logic it's a bug, and re-verify the dated tables before quoting a number.
+
+**Hand off instead when:** a new non-trivial feature has no approved spec + plan under `02-DOCS/wiki/sdd/` → stop and run [`specify`](../specify/SKILL.md) first (method: [`sdd`](../sdd/SKILL.md)), which routes back here once the plan is approved; one-line/low-risk changes go straight through. Anthropic-SDK internals (caching, thinking, batch) in a file that *only* imports `anthropic` → `claude-api` if your environment has it, since this skill stays multi-provider. Workspace scaffolding → [`harness`](../harness/SKILL.md). Choosing *which* coding agent to use → agent-eval territory. Pure prompt-wording tuning with no architecture change → prompt engineering, not this. A one-shot throwaway prompt, or no retrieval/tools/loop/evals at all → you don't need an agent; call the SDK directly and say so.
+
+## Decision rules (read before writing code)
+
+1. **Adapter first** — define the `LLMProvider` Protocol before any provider call.
+2. **Smallest loop that works** — single-agent before multi-agent; ReAct only when the path is uncertain; plan-execute when steps are knowable. Multi-agent means orchestrator-worker with a semaphore-bounded parallel fan-out, never a free-for-all.
+3. **Tools are typed contracts** — schema + validation + idempotency key on every side-effecting tool; no catch-all tools.
+4. **Retrieve, don't stuff** — RAG when ground truth lives in data; cite or refuse.
+5. **Eval before ship** — a golden set + regression gate in CI, or it's not production.
+6. **Cheapest model that passes the eval** — route/cascade up, never default to flagship.
+
+## The provider adapter (the heart of the skill)
+
+The one payload to internalize. Python 3.12+, Pydantic v2, **async** so it composes directly with the bounded loop (and orchestrator-worker fan-out) in [`references/agent-loops-and-harness.md`](references/agent-loops-and-harness.md). Structured output is the quirk that differs most per vendor: strict JSON Schema (OpenAI), tool-forcing (Anthropic), `response_json_schema` (Gemini). Streaming, the Gemini and OSS/litellm adapters, tool-result plumbing, and a `route()` registry live in [`references/provider-abstraction.md`](references/provider-abstraction.md) — this excerpt is the load-bearing core, not the whole interface.
+
+```python
+from __future__ import annotations
+
+import os
+from typing import Literal, Protocol, runtime_checkable
+
+from pydantic import BaseModel, Field
+
+
+class Message(BaseModel):
+    role: Literal["system", "user", "assistant", "tool"]
+    content: str
+
+
+class ToolSpec(BaseModel):
+    name: str
+    description: str
+    parameters: dict  # JSON Schema for the tool's arguments
+
+
+class Usage(BaseModel):
+    input_tokens: int = 0
+    output_tokens: int = 0
+    cost_usd: float = 0.0
+
+
+class CompletionRequest(BaseModel):
+    model: str  # resolved from config, e.g. "claude-sonnet-4-6" — never literal in logic
+    messages: list[Message]
+    tools: list[ToolSpec] = Field(default_factory=list)
+    response_schema: dict | None = None  # JSON Schema -> structured output
+    temperature: float = 0.0
+    max_tokens: int = 1024
+
+
+class CompletionResponse(BaseModel):
+    text: str = ""
+    tool_calls: list[dict] = Field(default_factory=list)  # [{id, name, arguments}]
+    usage: Usage = Field(default_factory=Usage)
+    raw: dict | None = None
+
+
+@runtime_checkable
+class LLMProvider(Protocol):
+    # Async so it drives the async agent loop directly. The full interface in
+    # references/provider-abstraction.md adds stream() and embed().
+    async def complete(self, req: CompletionRequest) -> CompletionResponse: ...
+
+
+class OpenAIAdapter:
+    def __init__(self, model: str) -> None:
+        from openai import AsyncOpenAI
+
+        self.model, self.client = model, AsyncOpenAI()
+
+    async def complete(self, req: CompletionRequest) -> CompletionResponse:
+        # Chat Completions shape (universal, still current); references/provider-abstraction.md
+        # gives the preferred Responses-API adapter. system stays a `system` role message here.
+        kwargs: dict = {"model": self.model, "messages": [m.model_dump() for m in req.messages],
+                        "temperature": req.temperature, "max_tokens": req.max_tokens}
+        if req.tools:
+            kwargs["tools"] = [{"type": "function", "function": {"name": t.name, "description": t.description, "parameters": t.parameters}} for t in req.tools]
+        if req.response_schema:
+            kwargs["response_format"] = {"type": "json_schema", "json_schema": {"name": "out", "schema": req.response_schema, "strict": True}}
+        r = await self.client.chat.completions.create(**kwargs)
+        msg = r.choices[0].message
+        calls = [{"id": c.id, "name": c.function.name, "arguments": c.function.arguments} for c in (msg.tool_calls or [])]
+        return CompletionResponse(text=msg.content or "", tool_calls=calls, raw=r.model_dump(),
+            usage=Usage(input_tokens=r.usage.prompt_tokens, output_tokens=r.usage.completion_tokens))
+
+
+class AnthropicAdapter:
+    def __init__(self, model: str) -> None:
+        from anthropic import AsyncAnthropic
+
+        self.model, self.client = model, AsyncAnthropic()
+
+    async def complete(self, req: CompletionRequest) -> CompletionResponse:
+        # QUIRKS: system is a top-level param (not a message); tools use input_schema (not function).
+        system = "\n".join(m.content for m in req.messages if m.role == "system") or None
+        turns = [{"role": m.role, "content": m.content} for m in req.messages if m.role != "system"]
+        kwargs: dict = {"model": self.model, "system": system, "messages": turns, "max_tokens": req.max_tokens, "temperature": req.temperature}
+        if req.tools:
+            kwargs["tools"] = [{"name": t.name, "description": t.description, "input_schema": t.parameters} for t in req.tools]
+        if req.response_schema:  # structured output via tool-forcing
+            kwargs["tools"] = [{"name": "out", "description": "Emit the result", "input_schema": req.response_schema}]
+            kwargs["tool_choice"] = {"type": "tool", "name": "out"}
+        r = await self.client.messages.create(**kwargs)
+        text = "".join(b.text for b in r.content if b.type == "text")
+        calls = [{"id": b.id, "name": b.name, "arguments": b.input} for b in r.content if b.type == "tool_use"]
+        return CompletionResponse(text=text, tool_calls=calls, raw=r.model_dump(),
+            usage=Usage(input_tokens=r.usage.input_tokens, output_tokens=r.usage.output_tokens))
+
+
+def get_provider(spec: str | None = None) -> LLMProvider:
+    """Parse 'provider:model' (default from env LLM) into a concrete adapter."""
+    provider, _, model = (spec or os.environ["LLM"]).partition(":")
+    if provider == "openai":
+        return OpenAIAdapter(model)
+    if provider == "anthropic":
+        return AnthropicAdapter(model)
+    raise ValueError(f"unknown provider: {provider!r}")
+# Gemini + OSS/litellm adapters, streaming, tool-result plumbing, and route() registry
+# -> references/provider-abstraction.md
+```
+
+## Good vs Bad
+
+Call-sites use the adapter and never name a model: `provider = get_provider(settings.llm)` (e.g. `"anthropic:claude-sonnet-4-6"`), then `await provider.complete(req)`. The two failures that survive that discipline:
+
+```python
+# BAD — parse-and-pray; wrong shape fails silently at 3am.
+raw = (await provider.complete(req)).text
+try:
+    data = json.loads(raw)
+except json.JSONDecodeError:
+    data = {}  # the bug is now invisible
+```
+
+```python
+# GOOD — strict structured output + schema validation that fails loudly on drift.
+class Answer(BaseModel):
+    sentiment: Literal["pos", "neg", "neu"]
+    score: float
+
+req.response_schema = Answer.model_json_schema()
+ans = Answer.model_validate_json((await provider.complete(req)).text)
+```
+
+```python
+# BAD — unbounded loop; no cap/timeout/idempotency. Burns budget, repeats side effects, wedges.
+while True:
+    resp = await provider.complete(req)
+    if not resp.tool_calls:
+        break
+    for call in resp.tool_calls:
+        await run_tool(call)
+```
+
+```python
+# GOOD — bounded loop: step cap + per-tool timeout + idempotency key (safe to retry).
+for step in range(max_steps):
+    resp = await provider.complete(req)
+    if not resp.tool_calls:
+        break
+    for call in resp.tool_calls:
+        async with asyncio.timeout(tool_timeout_s):
+            await run_tool(call, idempotency_key=call["id"])
+# full loop, budgets, recovery -> references/agent-loops-and-harness.md
+```
+
+## Tools & structured output (minimum viable)
+
+```python
+from typing import Callable, Literal
+
+from pydantic import BaseModel, ConfigDict, Field, ValidationError
+
+
+class CreateInvoiceArgs(BaseModel):
+    model_config = ConfigDict(extra="forbid")  # reject unknown keys from the model
+    customer_id: str = Field(min_length=1)
+    amount_cents: int = Field(gt=0)
+    currency: Literal["EUR", "USD"] = "EUR"
+
+
+class ToolResult(BaseModel):
+    status: Literal["success", "warning", "error"]
+    summary: str
+    data: dict | None = None
+    next_actions: list[str] = Field(default_factory=list)
+
+
+def _create_invoice(args: CreateInvoiceArgs) -> ToolResult:
+    invoice_id = f"inv_{args.customer_id}_{args.amount_cents}"  # real impl: DB insert + idempotency
+    return ToolResult(status="success", summary=f"Created {invoice_id}", data={"id": invoice_id})
+
+
+TOOLS: dict[str, tuple[type[BaseModel], Callable]] = {
+    "create_invoice": (CreateInvoiceArgs, _create_invoice),
+}
+
+
+def dispatch(name: str, raw_args: dict) -> ToolResult:
+    spec = TOOLS.get(name)
+    if spec is None:
+        return ToolResult(status="error", summary=f"unknown tool {name!r}", next_actions=["pick a registered tool"])
+    args_model, handler = spec
+    try:
+        args = args_model.model_validate(raw_args)  # validate BEFORE side effects
+    except ValidationError as e:
+        return ToolResult(status="error", summary="invalid args", data={"errors": e.errors()},
+                          next_actions=["fix the arguments and retry"])
+    return handler(args)
+```
+
+Schema design, sandboxing, idempotency, DI-scoped DB sessions, plus the RAG internals below — chunking, hybrid RRF, rerank, the citation grader, memory — are in [`references/tools-and-rag.md`](references/tools-and-rag.md).
+
+## RAG in 30 lines (provider-agnostic embeddings)
+
+```sql
+CREATE EXTENSION IF NOT EXISTS vector;
+CREATE TABLE IF NOT EXISTS docs (
+    id        bigserial PRIMARY KEY,
+    content   text NOT NULL,
+    embedding vector(1536) NOT NULL,
+    meta      jsonb NOT NULL DEFAULT '{}'
+);
+CREATE INDEX IF NOT EXISTS docs_embedding_hnsw
+    ON docs USING hnsw (embedding vector_cosine_ops);
+```
+
+```python
+async def embed(texts: list[str]) -> list[list[float]]:
+    # Same provider interface as completions; impl in references/tools-and-rag.md.
+    return await provider.embed(texts)  # returns one 1536-d vector per text
+
+
+async def retrieve(query: str, k: int = 5, min_sim: float = 0.25) -> list[dict]:
+    [q] = await embed([query])
+    rows = await db.fetch(  # cosine distance <=>; similarity = 1 - distance
+        "SELECT id, content, 1 - (embedding <=> $1) AS sim "
+        "FROM docs ORDER BY embedding <=> $1 LIMIT $2",
+        q, k,
+    )
+    return [dict(r) for r in rows if r["sim"] >= min_sim]
+
+
+async def answer(query: str) -> str:
+    chunks = await retrieve(query)
+    if not chunks:                                   # refuse rather than hallucinate
+        return "I don't have grounded information to answer that."
+    context = "\n".join(f"[{c['id']}] {c['content']}" for c in chunks)
+    req = CompletionRequest(
+        model=settings.model_id,
+        messages=[Message(role="system", content="Answer ONLY from context; cite chunk ids like [12]."),
+                  Message(role="user", content=f"{context}\n\nQ: {query}")],
+    )
+    return (await provider.complete(req)).text
+```
+
+## Evals & cost gates (the production line)
+
+```python
+import json
+import statistics
+import sys
+import time
+
+
+async def run_eval(golden_path: str, graders: list, thresholds: dict[str, float]) -> None:
+    cases = [json.loads(line) for line in open(golden_path)]  # {"input","expected","meta"}
+    results = []
+    for case in cases:
+        t0 = time.perf_counter()
+        out = await provider.complete(CompletionRequest(model=settings.model_id,
+              messages=[Message(role="user", content=case["input"])]))
+        scores = {g.name: g.grade(case, out) for g in graders}  # exact / schema / LLM-judge
+        results.append({"scores": scores, "cost": out.usage.cost_usd,
+                        "ms": (time.perf_counter() - t0) * 1000})
+    n = len(results)
+    metrics = {
+        "accuracy": sum(r["scores"]["exact"] for r in results) / n,
+        "faithfulness": sum(r["scores"]["judge"] for r in results) / n,
+        "p95_latency_ms": statistics.quantiles([r["ms"] for r in results], n=20)[-1],
+        "cost_per_task": sum(r["cost"] for r in results) / n,
+    }
+    failed = [k for k, lo in thresholds.items() if metrics[k] < lo]
+    print(json.dumps(metrics, indent=2))
+    sys.exit(1 if failed else 0)  # CI gate: non-zero blocks the merge
+```
+
+Routing cascade in one line: `route(task) → cheapest model whose eval passes; escalate only on a failed self-check`.
+Full runner, judge, CI gate, caching, batching and budgets →
+[`references/evals-and-observability.md`](references/evals-and-observability.md).
+
+## Observability (OTel GenAI, vendor-neutral)
+
+```python
+from opentelemetry import trace
+
+tracer = trace.get_tracer("agent")
+
+
+async def traced_complete(provider: LLMProvider, req: CompletionRequest) -> CompletionResponse:
+    with tracer.start_as_current_span("chat") as span:
+        span.set_attribute("gen_ai.system", settings.llm.split(":")[0])
+        span.set_attribute("gen_ai.request.model", req.model)
+        resp = await provider.complete(req)
+        span.set_attributes({"gen_ai.usage.input_tokens": resp.usage.input_tokens,
+                             "gen_ai.usage.output_tokens": resp.usage.output_tokens,
+                             "gen_ai.usage.cost_usd": resp.usage.cost_usd})
+        return resp
+# Langfuse / Phoenix / Braintrust are swappable OTLP backends: emit spans, swap the exporter.
+# span-per-tool, trace-id propagation, exporters -> references/evals-and-observability.md
+```
+
+## MCP: when and the smallest server
+
+**Native tools** when the agent and tools share a process/repo. **MCP** when tools must be reused across clients/teams or run out-of-process — accept the MCP cost (schema tokens, transport, ops) in exchange for reuse. TypeScript server, transports, HTTP+auth and testing are in [`references/mcp-servers.md`](references/mcp-servers.md).
+
+```python
+from fastmcp import FastMCP  # standalone fastmcp 2.x
+
+mcp = FastMCP("invoices")
+
+
+@mcp.tool()
+def create_invoice(customer_id: str, amount_cents: int, currency: str = "EUR") -> dict:
+    """Create an invoice. amount_cents must be > 0."""
+    if amount_cents <= 0:
+        raise ValueError("amount_cents must be positive")
+    return {"id": f"inv_{customer_id}_{amount_cents}", "currency": currency}
+
+
+@mcp.resource("invoice://{invoice_id}")
+def read_invoice(invoice_id: str) -> str:
+    """Read-only invoice lookup by id."""
+    return f"Invoice {invoice_id}: status=open"
+
+
+if __name__ == "__main__":
+    mcp.run()  # stdio transport
+# (MCP spec 2025-11-25; stateless-core RC 2026-07-28; verify before quoting)
+```
+
+## Anti-patterns
+
+| Anti-pattern | Reality |
+|---|---|
+| "I'll just call the OpenAI SDK directly, we'll never switch" | The adapter is ~40 lines; retrofitting it across 30 call-sites later is a rewrite. Adapter first. |
+| "JSON output is usually valid, I'll parse it" | "Usually" = pages at 3am. Use strict structured output + schema validation. |
+| "The agent loop works, I don't need a step cap" | Unbounded loops burn budget and wedge on errors. Cap steps, timeouts, and budget. |
+| "One mega-tool that takes a freeform command is flexible" | It's unobservable and unsafe. Narrow typed tools with idempotency keys. |
+| "We can eval by eyeballing outputs" | Vibes don't gate CI. Golden set + graders + threshold or it's not production. |
+| "Default everything to the flagship model, it's smartest" | 5–20× cost for no measured gain. Route to the cheapest model that passes the eval. |
+| "Stuff the whole doc in the prompt instead of RAG" | Blows context + cost and still hallucinates. Retrieve + cite + refuse. |
+| "Retry on every exception" | Retrying a 400/401 wastes budget. Retry only transient (429/5xx/timeout) with backoff+jitter. |
+| "Hardcode the model name, it's fine" | Names rot (Opus 4.7 → 4.8 in weeks). Resolve from config/registry. |
+| "MCP for everything" | In-process native tools are simpler and faster when reuse isn't needed. MCP only for cross-client reuse. |
+| "Tool results just return the raw API blob" | Give the model `status/summary/next_actions`; raw blobs waste context and stall recovery. |
+| "Prompt caching is Anthropic-only so skip caching" | Each provider has its own caching/dedup; abstract it behind the adapter, don't skip it. |
+
+## verify.sh
+
+`scripts/verify.sh` lints example agent code and dry-runs the eval smoke test in **the user's project** — not in this skill repo. It detects each tool (`ruff`, `mypy`, `tsc`/`node`, `go`, the eval entrypoint, `markdownlint`) and skips any that are missing with a yellow WARN; a missing tool never fails the run. Invoke it with `bash scripts/verify.sh` from the project root. Exit 0 means clean (or only skips); a non-zero exit means a real lint/typecheck/vet/eval failure.
+
+## Project grounding
+
+In a project with a `02-DOCS/` layer ([`harness`](../harness/SKILL.md)), read `02-DOCS/wiki/stack/agents.md` first on every use and stay consistent with it. If it is missing or stale, write this project's real choices there — provider(s) and model routing, where the adapter lives, tool/RAG conventions, eval gates, observability backend — as a `type: stack` article per the harness [`wiki-article-template.md`](../harness/references/wiki-article-template.md), index it in `02-DOCS/wiki/index.md`, and bump its `timestamp` in the same change as any convention change. No `02-DOCS/`? Skip silently — technical conventions here are *recorded, not gated*; never block the task on this.
+
+## See also
+
+- Stacks the examples target: [`fastapi`](../fastapi/SKILL.md), [`nextjs`](../nextjs/SKILL.md), [`go`](../go/SKILL.md), [`postgresdb`](../postgresdb/SKILL.md), [`flutter`](../flutter/SKILL.md). Harden with [`secure-coding`](../secure-coding/SKILL.md), ship with [`deployment`](../deployment/SKILL.md).
+- External (no sibling here; use if your environment provides them): `claude-api` for Anthropic-SDK-only tuning, `deep-research` for the research-harness fan-out / verify pattern.

package/skills/shared/building-agents/evals/README.md

@@ -0,0 +1,68 @@
+# Eval harness — `building-agents`
+
+This evaluates two things: (1) **triggering** — does the skill activate on the
+right prompts and stay quiet on near-misses — and (2) **capability** — does
+loading the skill measurably improve the agent's answer. These run through an
+**agent harness** (a Claude Code / SDK session), not a pure shell script: a
+human or a driver script feeds prompts and grades outcomes against
+`cases.yaml`.
+
+## Files
+
+- `cases.yaml` — `should_trigger`, `should_not_trigger`, and `capability` cases.
+
+## 1. Triggering eval
+
+Goal: confirm the skill's `description` routes correctly.
+
+1. Load **only** this skill into the agent (its `SKILL.md` description in the
+   skill index). For near-miss realism, also expose the sibling skills named in
+   each `route_to` so the agent can pick the better match instead of defaulting.
+2. For each prompt in `should_trigger` and `should_not_trigger`, start a fresh
+   session and send the prompt verbatim. Do **not** answer the task — observe
+   only whether `building-agents` is selected.
+3. Run **3–5 trials** per prompt (the selection is stochastic). Record the
+   fire-rate per prompt.
+4. Score:
+   - `should_trigger` item passes if the skill fires in a strong majority of
+     trials (>= 4/5).
+   - `should_not_trigger` item passes if the skill does **not** fire, and
+     ideally the agent routes to the `route_to` sibling (or to nothing when
+     `route_to: none`).
+
+**Pass bar:** >= 90% trigger accuracy across all items
+(true-positives + true-negatives / total trials). Any `should_not_trigger`
+that fires the skill is a false-positive and must be investigated — usually a
+too-greedy `description`.
+
+## 2. Capability eval
+
+Goal: prove the skill changes the output, not just the routing.
+
+1. For each `capability` scenario, run it **twice** in matched sessions:
+   - **WITHOUT** the skill loaded (baseline).
+   - **WITH** the skill loaded.
+2. Grade each answer against that scenario's `must_include` rubric: count how
+   many points are concretely present (code or prose), not merely gestured at.
+3. Compute coverage = points covered / total points, for both runs.
+
+**Pass bar:** WITH-skill coverage >= 80% of the rubric, AND a clear lift over
+the WITHOUT-skill baseline (the skill must add the load-bearing points the
+baseline misses — typically the adapter/Protocol, idempotency, cite-or-refuse,
+CI gate exit code, and transient-only retries). If the baseline already scores
+~as high, the skill isn't earning its place on those points.
+
+## Honest caveats
+
+- Selection is non-deterministic; always use multiple trials and report rates,
+  not a single pass/fail.
+- Grading `must_include` and "routes to the right sibling" requires judgment —
+  an LLM-as-judge can assist but spot-check it.
+- Keep `cases.yaml` in sync with `SKILL.md`: when the description's triggers or
+  the "Do NOT use" list change, update the cases in the same commit.
+
+## Counts
+
+- should_trigger: 7
+- should_not_trigger: 7
+- capability: 2

package/skills/shared/building-agents/evals/cases.yaml

@@ -0,0 +1,60 @@
+skill: building-agents
+
+should_trigger:
+  - prompt: "We're hardwired to the OpenAI SDK everywhere and the boss wants us to be able to switch to Claude or Gemini without a rewrite. How should I restructure this?"
+    why: "Provider-agnostic restructuring / 'swap models is a config change' is the core thesis of the skill (the capability interface + adapter)."
+  - prompt: "Add function calling to our support assistant so it can look up orders and issue refunds, but I don't want it to blow up if the model sends garbage arguments."
+    why: "Adding tool calling with schema validation + typed contracts + idempotency on side-effecting tools is squarely in scope."
+  - prompt: "Build me a chatbot that answers questions strictly from our internal handbook PDFs and says 'I don't know' when the answer isn't there."
+    why: "RAG over a private corpus with cite-or-refuse grounding — a named trigger — even though the user never says 'RAG', 'embeddings', or 'retrieval'."
+  - prompt: "Our agent sometimes spins forever and runs up a huge API bill. Help me make the loop safe and put a cost ceiling on each task."
+    why: "Bounded agent loop with step cap, timeouts, and token/cost budgets — the loop + cost-control trigger — phrased as a symptom, not a tool name."
+  - prompt: "Before we ship this LLM feature, I want a quality gate in CI that fails the build if accuracy or faithfulness drops on our test set."
+    why: "Golden set + graders + regression gate as a CI exit code is the 'eval before ship' decision rule and an explicit trigger."
+  - prompt: "I want to expose our internal pricing tools to several different AI clients so multiple teams can reuse them. What's the right way to wrap them?"
+    why: "Cross-client tool reuse out-of-process = build an MCP server, an explicit trigger; the user describes the reuse need rather than naming MCP."
+  - prompt: "Stand up tracing for our LLM calls so I can see token usage and cost per request in Langfuse without locking us to one backend."
+    why: "Vendor-neutral OTel GenAI observability with a swappable exporter is an explicit trigger."
+
+should_not_trigger:
+  - prompt: "Reword this prompt so the model gives more concise answers — no code changes, just better wording."
+    route_to: "none"
+    why: "Pure prompt-wording improvement with no architecture is prompt engineering; the skill's 'Do NOT use' list excludes it explicitly."
+  - prompt: "Write a one-off script that calls the OpenAI API once to summarize a string I paste in. No tools, no tests, throwaway."
+    route_to: "none"
+    why: "One-shot throwaway prompt with no tools/eval/production path — the skill says just call the SDK directly."
+  - prompt: "Tune Anthropic prompt caching and extended thinking in our anthropic-only client.py to cut latency on a file that only imports the anthropic SDK."
+    route_to: "external:claude-api"
+    why: "Single-vendor Anthropic-SDK internals tuning in a file that only imports anthropic defers to the dedicated Anthropic skill; this skill stays multi-provider."
+  - prompt: "Write the pgvector schema and an HNSW-indexed nearest-neighbour SQL query, plus tuning advice — just the database side, no LLM calls."
+    route_to: "postgresdb"
+    why: "Pure SQL / pgvector indexing with no embeddings pipeline, retrieval grounding, or agent is Postgres territory."
+  - prompt: "Set up the CI pipeline and Docker deploy for our FastAPI service that happens to host an LLM endpoint — focus on the deployment, not the model code."
+    route_to: "deployment"
+    why: "Shipping/CI/containerization of the service is deployment; the skill itself points to deployment for that concern."
+  - prompt: "Audit our auth handler for injection and secret-leak vulnerabilities — it's a normal REST endpoint, not an agent."
+    route_to: "secure-coding"
+    why: "A security-only review of a non-agent endpoint is secure-coding, not LLM agent architecture."
+  - prompt: "Bootstrap the docs/tools and docs/ workspace layout for this repo and generate the root AGENTS.md."
+    route_to: "harness"
+    why: "Workspace scaffolding (the docs/tools/docs/ layout) is explicitly delegated to harness."
+
+capability:
+  - scenario: "Our codebase calls `OpenAI().chat.completions.create(model='gpt-5.5', ...)` directly in three handlers. We need to support switching to Anthropic Claude or Gemini via config, with strict JSON output for a sentiment classifier and one side-effecting `create_invoice` tool. Lay out the architecture and the key code."
+    must_include:
+      - "Defines a capability/provider interface (e.g. an LLMProvider Protocol with a normalized request/response) that business logic programs against instead of a vendor SDK"
+      - "Model id is resolved from config/registry (e.g. 'provider:model'), never hardcoded in handler logic"
+      - "Per-provider adapters that hide vendor quirks (OpenAI tools/function vs Anthropic system-as-top-level-param + input_schema / tool-forcing for structured output)"
+      - "Strict structured output via JSON Schema with validation that fails loudly (e.g. Pydantic model_validate), not try/except json.loads that swallows errors"
+      - "The create_invoice tool has a typed args model with extra='forbid', validates arguments BEFORE side effects, and carries an idempotency key"
+      - "Returns a structured ToolResult (status/summary/next_actions), not the raw API blob"
+      - "Mentions selecting the cheapest model that passes an eval rather than defaulting to the flagship"
+  - scenario: "We have a RAG support bot and want to make it production-ready before launch. Show how to bound the agent loop, eval it, and observe it."
+    must_include:
+      - "Bounded agent loop: explicit step cap, per-tool timeout, and a token/cost budget (no while True)"
+      - "Retrieval grounded with cite-or-refuse: cite chunk ids and refuse when no chunk passes a similarity threshold rather than hallucinating"
+      - "A golden set plus graders (exact/schema/LLM-as-judge) with a CI threshold that exits non-zero to block the merge"
+      - "Reports metrics including accuracy/faithfulness, p95 latency, and cost per task"
+      - "OTel GenAI tracing with gen_ai.* attributes (model, input/output tokens, cost) and a swappable OTLP exporter/backend"
+      - "Retries only transient errors (429/5xx/timeout) with backoff+jitter, not 4xx like 400/401"
+      - "Routing/cascade to the cheapest model whose eval passes, escalating only on a failed self-check"