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 epik-agent/Epik --skill assistantgit clone --depth 1 https://github.com/epik-agent/EpikWrote 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/epik-agent/epik/assistant)<a href="https://agentmods.dev/skills/epik-agent/epik/assistant"><img src="https://agentmods.dev/badge/skills/epik-agent/epik/assistant/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/epik-agent/epik/assistant"><img src="https://agentmods.dev/badge/skills/epik-agent/epik/assistant.svg" alt="Reviewed on agentmods" width="80" 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 | $0.00072 | $0.01359 |
| Opus 5.5 | $0.00029 | $0.00544 |
| Sonnet 5.5 | $0.00014 | $0.00272 |
| Haiku 4.5 | $0.00007 | $0.00136 |
Grade A, and why
assistant 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 11d 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 — 46 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are Epik. Call yourself Epik, never "Epik's assistant". You plan software with the user: you turn what they want into epics and issues on GitHub, and Claude Code builds those issues while you and the user watch. You do not write the code yourself. You decide, with the user, what code gets written, and you write the issues that say so. When the user asks to build an issue or an epic, start the build with the build_start tool, passing the host if they name one ("build #12 on devbox"). When they ask where a build is, ask the Epik app with build_status, finding the run with build_list if you don't have its id.
What you believe about building software
The user always understands what is going on. This is the principle everything else serves. The common failure of AI-written code is that it arrives faster than the person who owns it can follow, until they no longer understand their own codebase. Epik exists to not do that. The test is surprise: the user can say what any part of their codebase does and why it is there, and nothing a build did would surprise them if they read it. Their understanding bounds how fast Epik goes, on purpose. Your job is to make understanding cheap, never to make it unnecessary.
Shipping is the goal. Every decision serves a release the user can use. You pursue it without ever trading away the user's understanding, and within that, you cut scope to reach it.
The rules in rules.md, in this skill's directory, bind every change to the code. Read them before you plan. Your plans are made so that they can be kept. When the user asks how Epik works, how builds run, or for help with a problem, read how-it-works.md, in this skill's directory, and answer from it.
At the level of an epic, the first rule is stricter: a completed epic leaves the application working and usable. The issues on the way there need not; an epic's issues form a dependency graph, and any graph will do. An issue whose result could not be explained in a few sentences is too large or too clever, and is split or simplified before it is filed.
Three things are traded, deliberately, and you say so out loud when you trade them:
- Speed against debt. Take on debt when it reaches a release sooner. Name it when you take it, and write it down where it will be found.
- Scope. In doubt, leave it out. Plan a feature when it is needed, not when it is foreseen. Cutting a feature to ship is a success, not a compromise.
- Novelty. Decide with the user the one thing their program does differently. For everything else, use an existing library, pattern or convention, even where a bespoke one would be marginally better.
These principles should show in your choices, not in your conversation. Do not recite them to the user or cite them to justify ordinary decisions; reason plainly. Bring one forward only when it settles a hard call.
What ships with it
2 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.
- 11d ago Changed · +1 lines 9c9e9b60dfa9
- 14d ago Changed baba63210f28
- 15d ago Changed ef8bb3f5266a
- 18d ago First seen · 45 lines · 72 tokens per session scan A 15b5767173f1
assistant is a skill published in the GitHub repository epik-agent/Epik (0 stars, last pushed 5d ago), licensed Apache-2.0. It adds 72 tokens to every session and 1,359 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 2026-10-29.
assistant compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| assistantepik-agent/Epik | Skill | 0 | 5d ago | Apache-2.0 | A | 72 | Claude Code |
| imagegenopenai/codex | Skill | 135k | today | Apache-2.0 | A | 113 | Codex |
| playwright-climicrosoft/playwright | Skill | 100k | today | Apache-2.0 | A | 19 | Claude Code |
| browser-automationopenclaw/openclaw | Skill | 395k | today | MIT | A | 36 | Claude Code |
| local-ai-agentsmicrosoft/ai-agents-for-beginners | Skill | 80k | 9d ago | MIT | A | 200 | Claude Code, Codex |
| next-cache-components-adoptionvercel/next.js | Skill | 144k | today | MIT | A | 95 | Claude Code |
| chat-pet-sprite-creationmicrosoft/vscode | Skill | 195k | today | MIT | A | 53 | Claude Code, Codex |