Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add skills/jeanibarz/knowledge-base-mcp-server/compare-embedding-modelsnpx skills add jeanibarz/knowledge-base-mcp-server --skill compare-embedding-modelsgit clone --depth 1 https://github.com/jeanibarz/knowledge-base-mcp-serverWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/jeanibarz/knowledge-base-mcp-server/compare-embedding-models)<a href="https://agentmods.dev/skills/jeanibarz/knowledge-base-mcp-server/compare-embedding-models"><img src="https://agentmods.dev/badge/skills/jeanibarz/knowledge-base-mcp-server/compare-embedding-models.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00045 | $0.01605 |
| Opus 5 | $0.00023 | $0.00803 |
| Sonnet 5 | $0.00009 | $0.00321 |
| Haiku 4.5 | $0.00005 | $0.00161 |
Grade A, and why
compare-embedding-models scanned grade A with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 6d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
| `OLLAMA_BASE_URL unreachable` | Ollama daemon not running | `ollama serve` in another terminal; verify with `curl $OLLAMA_BASE_URL/api/tags`. | How it starts
The opening of the file, as written. The whole thing — 102 lines — stays where its author put it; the contents beside it link to each section on GitHub.
When to use
- The user is choosing between two embedding models for a new knowledge base and wants concrete numbers (latency, cost, storage, quality) on their hardware, not a generic leaderboard.
- The user has switched models in the past and wants to verify the trade-off was worth it.
- The user is documenting a model choice for a team and needs an HTML artefact to attach to a decision record / RFC / PR.
Prerequisites
- knowledge-base-mcp-server
0.3.xinstalled (M0–M4 shipped; this skill needskb models {add, list}from §4.4 of RFC 013). - Both models reachable: Ollama running locally (
OLLAMA_BASE_URL) for ollama models;HUGGINGFACE_API_KEYset for HF models;OPENAI_API_KEYset for OpenAI models. The orchestrator reads provider tokens from env persrc/config.ts. - Disk space: ~10 MiB per model index for the medium synthetic fixture; more for larger profiles.
- For paid providers: estimated cost surfaced in the orchestrator preamble; non-zero requires
--yesor interactive confirmation (src/cli.tsrunAddModelflow, RFC 013 §4.4).
Steps
-
Identify model ids.
kb models listshows registered models with their<provider>__<slug>ids. If a target model is not yet registered, runkb models add <provider> <model_name>first (the orchestrator can auto-register with--yes, but registering explicitly lets the user audit cost upfront). -
Pick a fixture profile.
--fixture=small(~150 chunks) — sanity check, ~10 s.--fixture=medium(~600 chunks) — default; ~1 min on Ollama, ~30 s on HF/OpenAI.--fixture=external— runs against the corpus atKNOWLEDGE_BASES_ROOT_DIR(the user's real KB; no copy is made).--fixture=large(~3000-chunk arxiv corpus) — selection-grade. Not yet implemented in v1; deferred to M5.1 (RFC 013 §4.13.4 follow-up).
-
Run the comparison:
npm run bench:compare -- \ --models=ollama__nomic-embed-text-latest,huggingface__BAAI-bge-small-en-v1.5 \ --fixture=medium \ --concurrency=1,4,16
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 6d ago First seen · 102 lines · 45 tokens per session scan A 02f3f114ec95
compare-embedding-models is a skill published in the GitHub repository jeanibarz/knowledge-base-mcp-server (53 stars, last pushed 2d ago), licensed Unlicense. It adds 45 tokens to every session and 1,605 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
Inference
Guides model-serving and runtime-inference decisions across local, remote, and packaged deployment paths.
rag-perf
NVIDIA RAG Blueprint performance-tuning guidance for profiling retrieval stacks, comparing bottlenecks, and validating latency or throughput improvements.
arrowspace
Spectral vector search using graph Laplacian eigenstructure. Use when cosine/L2 similarity misses latent structure in your embeddings.
hf-gradio
Use Gradio applications via API. Use when the user asks for to generate a prediction from a Gradio app on Hugging Face spaces or public URL. For example, "Generate an image using black-forest-labs/FLUX.2-dev".
sentence-transformers
Framework for state-of-the-art sentence, text, and image embeddings. Provides 5000+ pre-trained models for semantic similarity, clustering, and retrieval. Supports multilingual, domain-specific, and multimodal models. Use for generating embeddings for RAG, semantic search, or similarity tasks. Best for production…
together-ai-inference
Serverless inference, fine-tuning, embeddings, image generation, and batch processing on 200+ open-source models via an OpenAI-compatible API. Use when you need fast, cost-effective access to open-source LLMs without managing infrastructure.