Gdd Bandit Reset logo

Gdd Bandit Reset

Community
hegemonart
gdd-bandit-reset

Confirm-then-reset the per-(agent, bin, delegate) bandit posterior - backs up .design/telemetry/posterior.json to posterior.json.bak, then clears it to a fresh empty envelope. Mutation companion to read-only bandit-status. Use when the posterior is corrupted/unparseable, after a major agent/skill roster change invalidates accumulated arms, or when you deliberately want to rebootstrap adaptive routing from informed priors.

Overview

Publisherhegemonart
Repositoryget-design-done
Skill namegdd-bandit-reset
Stars
3
Forks
1
Bundled files
Instructions only
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

    Published by hegemonart on GitHub. Read the source before you install it.

Installation

Install the Gdd Bandit Reset AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/hegemonart/get-design-done/tree/main/scripts/skill-templates/bandit-reset
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/hegemonart/get-design-done.git /tmp/get-design-done
mkdir -p .claude/skills
cp -r /tmp/get-design-done/scripts/skill-templates/bandit-reset .claude/skills/gdd-bandit-reset
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Gdd Bandit Reset in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Gdd Bandit Reset on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Gdd Bandit Reset is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

gdd-bandit-reset

Role

You are a deterministic, destructive maintenance skill. You are the ONLY skill that clears the bandit posterior - the mutation companion to read-only {{command_prefix}}bandit-status. You read the posterior path declared by scripts/lib/bandit-router.cjs's DEFAULT_POSTERIOR_PATH (.design/telemetry/posterior.json), REQUIRE explicit confirmation, back the file up to posterior.json.bak, then overwrite it with a fresh empty envelope so the next bandit pull rebootstraps from informed priors. See ./reference/bandit-integration.md for setup, interpretation, and convergence guidance.

Invocation Contract

  • Input: optional --yes to skip the interactive confirmation (for non-interactive/automated runs).
  • Output: a Markdown reset receipt to stdout (backup path + arms cleared + envelope written).

Procedure

1. Locate the posterior file

Read .design/telemetry/posterior.json (path declared by scripts/lib/bandit-router.cjs's DEFAULT_POSTERIOR_PATH - never hardcode a different path). Missing → nothing to reset; emit and skip to Section 5 (Record):

## Bandit Posterior Reset

No posterior file found at `.design/telemetry/posterior.json` — nothing to reset.

The next bandit pull with `adaptive_mode: full` will bootstrap a fresh posterior from informed priors. See `reference/bandit-integration.md`.

Note: the posterior only learns (updates from outcomes) on the SDK / headless session-runner path. In interactive Claude Code with adaptive_mode: full, the bandit samples from the configured priors but does not currently update them in-session. A reset therefore re-bootstraps the priors the SDK path will subsequently learn from. See reference/bandit-integration.md ("Where adaptive routing actually learns").

If present, count the arms (arms.length, treating a missing/non-array arms as 0) so the confirmation and receipt can report what will be cleared. A corrupted/unparseable file is still resettable - report arms: unknown (file unparseable) and continue.

2. Require explicit confirmation

This is a DESTRUCTIVE operation. Do NOT proceed without confirmation.

  • If --yes was passed, skip straight to Section 3.
  • Otherwise, prompt via AskUserQuestion: "Reset the bandit posterior at .design/telemetry/posterior.json? This clears learned arms. A backup will be written to posterior.json.bak first." with options Reset and Cancel.
  • On Cancel (or any non-affirmative answer), abort WITHOUT touching either the posterior or the backup, and emit:
## Bandit Posterior Reset — Cancelled

No changes made. The posterior at `.design/telemetry/posterior.json` is untouched (<N> arms).

Then skip to Section 5 (Record) with reset: false.

3. Back up the current posterior

Copy the live posterior to .design/telemetry/posterior.json.bak (sibling backup) BEFORE clearing it, so the previous state is always recoverable. Overwrite any existing .bak from a prior reset. If the backup write fails, ABORT before clearing; never clear without a successful backup.

4. Clear to a fresh empty envelope

Overwrite .design/telemetry/posterior.json with a fresh empty envelope matching scripts/lib/bandit-router.cjs's loadPosterior() shape (SCHEMA_VERSION = 1.0.0, current ISO generated_at, empty arms):

json
{
  "schema_version": "1.0.0",
  "generated_at": "<ISO>",
  "arms": []
}

Write atomically where possible (.tmp + rename, mirroring savePosterior()). Then emit the receipt:

## Bandit Posterior Reset

Posterior cleared. The next bandit pull with `adaptive_mode: full` will rebootstrap from informed priors.

- Backup: `.design/telemetry/posterior.json.bak`
- Arms cleared: <N>
- Fresh envelope: `.design/telemetry/posterior.json` (schema_version 1.0.0, 0 arms)

Restore the previous state with: `cp .design/telemetry/posterior.json.bak .design/telemetry/posterior.json`
Verify the cleared state with `{{command_prefix}}bandit-status`. See `reference/bandit-integration.md`.

5. Record

Append one JSONL line to .design/skill-records.jsonl: {"skill":"gdd-bandit-reset","ts":"<ISO>","reset":<bool>,"arms_cleared":<count>,"backup_written":<bool>}. The skill mutates ONLY the posterior (+ its .bak) and appends to skill-records.jsonl (telemetry); it touches no other state.

Cross-references

  • {{command_prefix}}bandit-status - read-only companion; inspect the posterior before/after a reset.
  • ./reference/bandit-integration.md - operator guide; interpretation patterns and when a reset is warranted.
  • scripts/lib/bandit-router.cjs - posterior shape, DEFAULT_POSTERIOR_PATH, SCHEMA_VERSION, loadPosterior(), savePosterior(), reset().

Frequently asked questions

What does the Gdd Bandit Reset AI skill do?

Confirm-then-reset the per-(agent, bin, delegate) bandit posterior - backs up .design/telemetry/posterior.json to posterior.json.bak, then clears it to a fresh empty envelope. Mutation companion to read-only bandit-status. Use when the posterior is corrupted/unparseable, after a major agent/skill roster change invalidates accumulated arms, or when you deliberately want to rebootstrap adaptive routing from informed priors.

Why use Gdd Bandit Reset on TypingMind?

Because you install it once and use it with any model. Gdd Bandit Reset is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Gdd Bandit Reset in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/hegemonart/get-design-done/tree/main/scripts/skill-templates/bandit-reset. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Gdd Bandit Reset?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Gdd Bandit Reset?

As many as you like. As long as a model supports skills, you can use Gdd Bandit Reset with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Gdd Bandit Reset AI skill free?

Yes. It is published on GitHub by hegemonart under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇