Borrowing it
Nothing to install: this file belongs to yantrikos/yantrikdb. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/yantrikos/yantrikdb/main/.claude/skills/build-pack/SKILL.mdgit clone --depth 1 https://github.com/yantrikos/yantrikdbWrote 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/yantrikos/yantrikdb/build-pack)<a href="https://agentmods.dev/skills/yantrikos/yantrikdb/build-pack"><img src="https://agentmods.dev/badge/skills/yantrikos/yantrikdb/build-pack/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/yantrikos/yantrikdb/build-pack"><img src="https://agentmods.dev/badge/skills/yantrikos/yantrikdb/build-pack.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00088 | $0.02321 |
| Opus 5 | $0.00044 | $0.01161 |
| Sonnet 5 | $0.00018 | $0.00464 |
| Haiku 4.5 | $0.00009 | $0.00232 |
Grade A, and why
build-pack scanned grade A with 0 findings 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 10d 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.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 209 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Building a pack
A pack is a sealed, signed YantrikDB file a model mounts to gain knowledge and behaviour it lacks, and unmounts without a trace. This skill covers authoring one that is worth shipping — which is decided by measurement, not by how good the corpus looks.
The one thing to get right first
A pack shifts knowledge and procedure. It does not raise raw capability. Every claim must be phrased so it can be checked:
- ✗ "makes your model write like Shakespeare"
- ✓ "raises compliance with Shakespearean craft rules from 7/12 to 10/12 on qwen3.5:4b, on checks written down before the run"
If you cannot state the claim as a measured delta on named checks, you do not yet know whether the pack works.
Which of the four shapes is this?
Measured across three model sizes; the shape predicts the payoff and the buyer.
| Shape | Example | Typical lift | Who buys it |
|---|---|---|---|
| Knowledge the model cannot have | private codebase, post-cutoff API | huge (1/20 → 18/20) | anyone running a local model |
| Method it knows but doesn't apply | reasoning discipline, review procedure | large at every size (+5 to +8) | everyone — the most underserved shape |
| Framework house rules | React, WordPress conventions | moderate; ceiling is high already | small-model operators |
| Style craft rules | period voice, brand tone | real but smallest | small-model operators |
Two findings worth carrying: method packs lift even a 27B, because models know what a procedure is and reach for it only when it is in front of them. Craft and domain packs are inverse to model size — a 27B often scores near ceiling unaided, so say who the pack is for instead of promising uniform lift.
The three tiers
pack/
pack.toml identity, namespace, ingest defaults, coverage
corpus.md the knowledge — retrieved on similarity
constitution.md the rules — injected on EVERY turn (optional)
eval.jsonl questions with deterministic expectations
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.
- 10d ago First seen · 209 lines · 88 tokens per session scan A 3204f3c74c04
build-pack is a skill published in the GitHub repository yantrikos/yantrikdb (58 stars, last pushed today), licensed Apache-2.0. It adds 88 tokens to every session and 2,321 once invoked, about $0.0004 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
evaluating-llms-harness
Evaluates LLMs across 60+ academic benchmarks (MMLU, HumanEval, GSM8K, TruthfulQA, HellaSwag). Use when benchmarking model quality, comparing models, reporting academic results, or tracking training progress. Industry standard used by EleutherAI, HuggingFace, and major labs. Supports HuggingFace, vLLM, APIs.
qdrant
Vector search engine for production RAG systems.
chroma
Embedding database for RAG and semantic search.
pinecone
Managed vector DB for production RAG and search.
cli-eval
Create and run evaluation suites, watch live benchmark progress, view scorecards, compare model performance, and integrate eval runs with CI workflows from the CLI.
model-merging
Merge multiple fine-tuned models using mergekit to combine capabilities without retraining. Use when creating specialized models by blending domain-specific expertise (math + coding + chat), improving performance beyond single models, or experimenting rapidly with model variants. Covers SLERP, TIES-Merging, DARE, Task…