Skip to content
Agentmart
Submit a tool

nk-chat

Load Niko's memory bank and begin a conversation

Skill by Texarkanine

IndexedAGPL-3.0
Write a review

Chalk Labs

Need help deploying nk-chat?

Chalk Labs sets up agents like this inside real teams: connected to your tools, tested on your data, with a human in the loop where it matters.

Book a call

Install

Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.

npx agentmarket install skill/texarkanine-nk-chat-d0d5d4

Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.

Get the weekly diff

The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.

We use your email only for this newsletter. Privacy

About

Load Niko's memory bank and begin a conversation

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.

Tools
No tools found
Secrets
None asked for
Install scripts
None found

Network hosts it talks to 1

  • codeload.github.com · found in install command

Files it touches

  • ~/.claude/skills/nk-chat

What's inside

Skill name
nk-chat
Description
Load Niko's memory bank and begin a conversation
License
AGPL-3.0
Instructions
69 lines in SKILL.md · view on GitHub

Files (1)

  • SKILL.md

SKILL.md (first lines)

## Step 1: Load Context

Read everything available, in this order. Do not skip files.

1. **Persistent files** (always required if memory bank exists):
   - `memory-bank/productContext.md`
   - `memory-bank/systemPatterns.md`
   - `memory-bank/techContext.md`

2. **Ephemeral files** (read if `memory-bank/active/` exists; do not create or modify):
   - `memory-bank/active/projectbrief.md`
   - `memory-bank/active/activeContext.md`
   - `memory-bank/active/tasks.md`
   - `memory-bank/active/progress.md`
   - any files under `memory-bank/active/creative/`
   - any files under `memory-bank/active/reflection/`

3. **Recent archive entries** (skim, do not exhaustively read): list `memory-bank/archive/` to know what past work exists. Read individual archive files only when a question makes one relevant.

### Graceful Degradation

- **No `memory-bank/` directory at all**: Stop loading. Inform the operator that no memory bank exists and offer to initialize one via `/niko`. Do not fabricate context. Do not continue to Step 2. You are done.
- **Persistent files missing or partial**: Inform the operator which files are missing. Offer to initialize via `/niko`. Proceed with whatever context is available, but be explicit about gaps when answering questions.
- **No `memory-bank/active/` directory**: Note "no task currently in flight" and proceed with persistent-context-only Q&A. This is normal and expected.
- **Partial ephemeral state** (e.g., `progress.md` exists but no `tasks.md`): Note the inconsistency factually ("looks like a task may be paused or in an unusual state"). Do not attempt to repair it.

## Step 2: Greet & Orient

Print a structured "Context Loaded" summary so the operator knows exactly what you are grounded in. Use this exact shape:

~~~markdown
# Context Loaded

**Persistent context:** [numbered list of persistent files read, or "none — no memory bank found"]

**In-flight task:** [one-line summary from `activeContext.md` and `progress.md`, or "none — no active ephemeral state"]

**Phase:** [current phase from `activeContext.md`, if any]

**Recent archives:** [numbered list of most-recent 1–3 by name, or "none"]
~~~

Then handle the operator's input:

- **No question provided alongside `/nk-chat`**: append `What would you like to discuss?` and wait for input.
- **Question provided alongside `/nk-chat`**: proceed directly to Step 3.

## Step 3: Conversational Q&A Loop

Answer questions using the loaded context as ground truth. Apply these guidelines:

- **Cite sources.** When an answer comes from a specific memory-bank file, archive entry, or codebase file, name it. The operator should be able to verify your grounding.
- **Acknowledge gaps honestly.** If the loaded context doesn't answer a question, say so. You may use read-only codebase exploration tools to fill the gap, but cite what you read.
- **Ask clarifying questions when needed.** Don't guess at intent on ambiguous questions.
- **Stay grounded.** Do not speculate about future work, suggested designs, or "what we should do" beyond what is supported by the loaded context. If the operator asks for an opinion, give one — but mark it clearly as opinion, not as project fact.

The conversation continues for as many turns as the operator wants. Each turn: read the question, answer using context, cite sources, repeat.

## Step 4: Handoff Triggers

If the conversation reveals real work to be done, **do not do the work**. Hand off explicitly:

- Operator asks you to make a change → "That's real work — you'll want to invoke `/niko` so the work goes through the proper workflow."
- Conversation surfaces an open design question worth exploring with a documented outcome → "Sounds like a `/niko-creative` candidate. Want me to summarize the question for you to take into that?"
- Operator says "yeah just do it" or otherwise tries to skip the handoff → restate the contract once: "Chat is read-only by design — even with your go-ahead I shouldn't modify state from here. Drop into `/niko` and I'll do it properly there." Do not capitulate.

You may help the operator *prepare* for a handoff (e.g., "here's what I'd put in the project brief if you ran `/niko` next") — that's still a read-only conversational artifact, not a state change.

From Texarkanine/nervouscsstem, AGPL-3.0.

More from Texarkanine/nervouscsstem

Signals

Stars
0
Forks
0
Open issues
8
Last commit (this folder)
Sep 12, 2026 (1 month ago)
Last release
Sep 13, 2026 (1 month ago)

From GitHub (fetched Oct 24, 2026).

Details

License
AGPL-3.0
First seen
Oct 24, 2026
Last checked
Oct 24, 2026

Same skill elsewhere

Copies of this SKILL.md published by others; search shows only one of them.

More from Texarkanine

Done for you

Rather not set it up yourself?

Chalk Labs scopes, builds and runs agents for your team: the task audit, the eval harness, the integrations and the weekly ops report. Fixed scope, priced before we start.

See how we build AI agents →Read the case studies

Chalk Labs is a two-founder AI and growth studio. We build AI agents, CRMs and MVPs for Web3 and AI companies, and have shipped 50+ product and token launches.

launches shipped
50+
launches shipped
media placements
500+
media placements
founders on every account
2
founders on every account

Press we've placed: Forbes · CoinDesk · Decrypt · Cointelegraph · The Block

Rahil Jain, Co-founder, Chalk Labs. Ex-Goldman Sachs engineer. You talk to the person who builds it.

Get nk-chat running in your business

Tell us what you want to automate. A founder replies within one working day with a plan, not a pitch deck.

Goes to the Chalk Labs founders, not the newsletter. Privacy

Need help deploying nk-chat?

CallGet help
Cite this page

nk-chat skill. Agent Mart. https://agentmart.dev/skills/texarkanine-nk-chat-d0d5d4 (last updated 2026-10-24)