side-quest

A command that sends a task to one or more background coding agents and reports the result in a Dungeons & Dragons style alongside a plain-English summary.

In plain words
What is it for?
Use it to delegate a self-contained task, such as an investigation, refactor, or other coding quest. Put the task in `$ARGUMENTS`; if it is empty, the command asks what to do.
Why use it?
It keeps the main conversation available while another agent handles the task. It also requires missing choices or approval for outward-facing and destructive actions before starting.

Command

Install

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.

agentmods
npx agentmods add commands/trtmn/agent-plugins/side-quest
Clone the repo
git clone --depth 1 https://github.com/trtmn/agent-plugins
Per session 0 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,006 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.00000 $0.01006
Opus 5 $0.00000 $0.00503
Sonnet 5 $0.00000 $0.00201
Haiku 4.5 $0.00000 $0.00101

Measured yesterday against content hash fe0496852cdd, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

side-quest 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 yesterday.

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.

plugins/side-quest/commands/side-quest.md · 86 lines

How it starts

The opening of the file, as written. The whole thing — 86 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Take the task in $ARGUMENTS and run it as a side quest: hand it off to one or more background subagents so this conversation stays free, then report in Dungeons & Dragons style — always paired with a plain-English summary.

1. Parse the quest

$ARGUMENTS is the task. If it's empty, ask in one line what quest the user seeks (tavern-keeper flavor + plain English) and stop until they answer.

If the quest needs user input to proceed (ambiguous target, missing credentials, a choice only the user can make), resolve that before launching — background agents cannot ask questions mid-quest.

If the quest is destructive or outward-facing (push, deploy, delete, send, publish), confirm with the user before launch.

2. Rate the Challenge Rating and pick the model

CR Difficulty Examples Model
1–2 Trivial lookups, renames, single-file tweaks 🧝 haiku
3–4 Standard refactors, multi-file features, research 🧙 sonnet
5+ Boss fight complex, cross-cutting, architectural 🐲 opus

If the user names a model anywhere in the quest, that always overrides the rating.

3. Assemble the party

  • Default: one general-purpose agent.
  • If the quest decomposes into independent pieces, launch a party — multiple agents in a single message so they run concurrently, each with a clear sub-quest.
  • Every agent launches with run_in_background: true and the chosen model.
  • Every agent prompt must be self-contained: working directory, constraints, definition of done, and an instruction to return raw results (their final message comes back to you, not the user).

4. Announce the quest

Respond with the full-immersion acceptance scroll, then the mandatory plain-English block:

📜 **QUEST ACCEPTED: <evocative quest name>**
⚔️ CR <n> (<difficulty>) — <party composition, e.g. "a lone 🧙 Sonnet wizard rides forth">
🎲 Initiative rolled… the party ventures into the background!

> **Plain English:** Launched <N> background agent(s) (<model>) to
> <what they're doing>. I'll report back when they finish.

Read the full file on GitHub · 86 lines

Changes

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.

  1. yesterday First seen · 86 lines · 0 tokens per session scan A fe0496852cdd

Subscribe to this mod's changes

side-quest is a command published in the GitHub repository trtmn/agent-plugins (2 stars, last pushed 4d ago), licensed Unlicense. It costs nothing until one of its globs matches a file; then it loads 1,006 tokens. 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 2026-08-31.