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 agentmods add agents/isaacsight/kernel/producer_engineer_specgit clone --depth 1 https://github.com/isaacsight/kernelWrote 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/agents/isaacsight/kernel/producer_engineer_spec)<a href="https://agentmods.dev/agents/isaacsight/kernel/producer_engineer_spec"><img src="https://agentmods.dev/badge/agents/isaacsight/kernel/producer_engineer_spec.svg" alt="Measured on agentmods" 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 | $0.00000 | $0.22081 |
| Opus 5 | $0.00000 | $0.11041 |
| Sonnet 5 | $0.00000 | $0.04416 |
| Haiku 4.5 | $0.00000 | $0.02208 |
Grade A, and why
PRODUCER_ENGINEER_SPEC 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.
How it starts
The opening of the file, as written. The whole thing — 2,196 lines — stays where its author put it; the contents beside it link to each section on GitHub.
K:BOT Producer + Sound Engineer Agent — Architecture Specification
Agent ID: producer (upgraded from current)
File: packages/kbot/src/agents/producer.ts
New companion: packages/kbot/src/tools/producer-engine.ts
Version: v2.0 — Unified Producer + Sound Engineer
1. Problem Statement
The current producer agent is a DAW controller. It knows how to send OSC commands and has a solid knowledge base, but it lacks two critical capabilities:
-
Producer intelligence — When a user says "make a trap beat," the current agent needs the user to specify every detail. A real producer would make 50+ creative decisions instantly: key, tempo, scale, instruments, patterns, voicings, arrangement, sound selection.
-
Sound engineer intelligence — After creating tracks, the current agent does not auto-mix. A real engineer would set volumes, pan positions, EQ, compression, sends, sidechain, and master chain without being asked.
This spec designs a unified agent that combines both. One prompt, one engine, one shot.
2. Architecture Overview
User: "make a trap beat"
│
▼
┌─────────────────────────┐
│ PRODUCER AGENT │ (system prompt in producer.ts)
│ Understands intent │
│ Picks genre │
└────────┬────────────────┘
│
▼
┌─────────────────────────┐
│ GENRE DECISION ENGINE │ (producer-engine.ts)
│ Resolves all creative │
│ decisions from genre │
│ preset + randomization│
└────────┬────────────────┘
│
▼
┌─────────────────────────┐
│ EXECUTION PIPELINE │ (producer-engine.ts)
│ 13-step sequence │
│ using existing tools │
└────────┬────────────────┘
│
▼
┌─────────────────────────┐
│ MIX ENGINE │ (producer-engine.ts)
│ Auto-applies volumes, │
│ panning, sends, EQ │
│ based on genre preset │
└────────┬────────────────┘
│
▼
Report to user
The engine is a new tool (produce_beat / produce_track) registered alongside the existing Ableton tools. The agent's system prompt instructs it to call this tool for full-track requests, while still having access to individual tools for surgical operations.
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.
- yesterday First seen · 2,196 lines · 0 tokens per session scan A 9f508e9fc178
PRODUCER_ENGINEER_SPEC is an agent published in the GitHub repository isaacsight/kernel (16 stars, last pushed 6d ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 22,081 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 2026-09-03.
Other agents, from other repositories
board-opus
Design-taste and architecture executor on the delivery board (opus tier). Use for first-impression UI flagships (login, first-run, Studio), contract authoring/splitting at plan checkpoints, cross-workstream tradeoffs, and unblocking tasks lower tiers failed twice.
board-sonnet
Standard delivery-board contract executor (sonnet tier). Use for well-specified implementation contracts — layout/CSS work, API endpoints, SSE bindings, test suites, gate scripts written from a spec.
board-verifier
Independent verifier for delivery-board tasks (haiku tier, fresh context). Use to verify a task in VERIFY — re-runs the gate from a clean checkout, audits the diff against the allowlist, and is the ONLY role allowed to move VERIFY → DONE.
board-haiku
Mechanical delivery-board executor (haiku tier). Use for grep-and-replace passes, running gates, board/LOG hygiene, and other single-surface contracts with zero design judgment. Also see board-verifier for verification work.
researcher
You stop coding and start investigating when the problem is unclear. Every problem can be solved with enough information.
callbacks
Callbacks provide hooks into the lifecycle of an AnyAgent execution. Using callbacks, you can monitor, control, and extend agent behavior without modifying the core underlying agent logic.