SkillPack packages AI skills and tools into runnable local agents that teams can operate in their own environment. Teams use it to keep data local while accessing agents through Slack or Telegram, and the catalogue add-ons are agents and skills built for its workflow.
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 skills add CreminiAI/skillpack --skill skillpack-creatorgit clone --depth 1 https://github.com/CreminiAI/skillpackWrote 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/creminiai/skillpack/skillpack-creator)<a href="https://agentmods.dev/skills/creminiai/skillpack/skillpack-creator"><img src="https://agentmods.dev/badge/skills/creminiai/skillpack/skillpack-creator/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/creminiai/skillpack/skillpack-creator"><img src="https://agentmods.dev/badge/skills/creminiai/skillpack/skillpack-creator.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 1 finding, up to medium
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- medium MCP Rug Pull · line 79 npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.Fix: Pin the version: npx @scope/[email protected]
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 | $0.00067 | $0.00931 |
| Opus 5.5 | $0.00027 | $0.00372 |
| Sonnet 5.5 | $0.00013 | $0.00186 |
| Haiku 4.5 | $0.00007 | $0.00093 |
Grade A, and why
skillpack-creator 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 1mo 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 — 100 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Skillpack Creator
Overview
Turn a successful task into a reusable SkillPack. Extract the stable workflow, decide what belongs in a local skill versus pack-level prompts, generate the pack structure, and package it only after the workflow is explicit and repeatable.
Workflow
1. Normalize the source task
Reduce the finished task into a clean execution spec:
- Capture the user goal, concrete deliverable, and the final successful workflow (not the full exploratory transcript).
- List required skills, tools, files, secrets, and environment assumptions.
- Separate deterministic steps from heuristic steps; remove dead ends and debugging noise.
- If the task is still too broad, narrow the scope instead of writing a vague mega-skill. If key success conditions depend on hidden human judgment, mark the pack as a best-effort assistant workflow.
Ask for missing stable facts or infer only the low-risk pieces.
2. Decide what the pack should contain
- Local skill (
skills/): reusable procedural knowledge. Keep scripts minimal unless reproducibility depends on exact file generation or repetitive shell steps. - Scripts (
scripts/): repeated shell or file-generation logic where reliability matters. - References (
references/): detailed schemas, API notes, or conventions that should not bloatSKILL.md. - Prompts (
skillpack.json): 1–3 pack-level starter inputs for the UI — not a DAG or state machine. Seereferences/skillpack-format.mdfor exact pack semantics.
3. Create the pack specification
Before writing files, define the pack spec. Prefer one local orchestrator skill plus a small number of external skills. Example minimal manifest:
{
"name": "company-research",
"description": "Research a company and produce a summary report",
"version": "1.0.0",
"prompts": ["Research {company} and create a report with financials and competitors"],
"skills": [
{ "name": "research-orchestrator", "source": "./skills/research-orchestrator", "description": "Orchestrate company research across multiple sources" }
]
}
What ships with it
3 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 1mo ago First seen · 100 lines · 67 tokens per session scan A bd4992fa65dc
skillpack-creator is a skill published in the GitHub repository CreminiAI/skillpack (1,204 stars, last pushed 1mo ago), licensed MIT. It adds 67 tokens to every session and 931 once invoked, about $0.0003 per session on Opus 5.5. 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
project-scaffolder
Use when a learning plan requires runnable chapter files, a multi-lesson coding course, a project-based route, or a workspace the learner will open in an editor.
adaptive-lesson-flow
Use when generating a structured learning deck, a lesson manifest, or a short practice sequence for a learner with a confirmed learning profile.
concept-teaching
A guided teaching workflow for learning one programming concept at a time in Python, Go, or the code you are working on. TDD means test-driven development, but this description does not say that the add-on teaches TDD.
learning-intent-router
Use when a learner expresses a new learning goal, interview or exam need, provides materials, corrects a prior goal, or answers an onboarding question.
learning-plan
A Chinese-language planning tool for creating or adjusting short, practical learning routes for Python, Go, mixed-language, or goal-focused programming study.
learner-onboarding
A learning-guidance skill for understanding a developer's goals, background, available time, and code project before creating a study plan. It also routes direct questions about individual concepts to a shorter learning path.