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 instructions/dictionlabs/diction/agents-mdgit clone --depth 1 https://github.com/DictionLabs/DictionWrote 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/instructions/dictionlabs/diction/agents-md)<a href="https://agentmods.dev/instructions/dictionlabs/diction/agents-md"><img src="https://agentmods.dev/badge/instructions/dictionlabs/diction/agents-md.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 | $0.03074 | $0.03074 |
| Opus 5 | $0.01537 | $0.01537 |
| Sonnet 5 | $0.00615 | $0.00615 |
| Haiku 4.5 | $0.00307 | $0.00307 |
Grade A, and why
Diction AGENTS.md 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 3d 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 — 325 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Diction Gateway -- Agent Reference
Machine-legible reference for the Diction self-hosted gateway. Covers routes, wire formats, environment variables, capabilities, and what the gateway can and cannot do.
Overview
The gateway is a Go HTTP service that sits between the Diction iOS app and one or more OpenAI-compatible speech-to-text backends. It handles:
- WebSocket streaming for low-latency transcription
- HTTP transcription fallback
- Optional LLM post-processing (transcript cleanup, voice editing, suggestions)
- Optional trial token auth (Diction One -- not for self-hosters)
WARNING: TEXT_ROUTES_OPEN=false (default). /v1/text/* routes return 403 until you explicitly set TEXT_ROUTES_OPEN=true or AUTH_ENABLED=true. This is a deliberate speed bump -- not a security control. Set it before calling those routes.
Routes
GET /health
Health check. Returns 200 when the gateway is up.
Response: 200 OK
Body: "ok"
GET /v1/models
Lists configured speech backends (OpenAI-compatible + Diction legacy grouping). Also includes a top-level capabilities object describing what this gateway instance can do (additive -- never removes data[] or providers[]).
Response: 200 OK
Content-Type: application/json
{
"object": "list",
"data": [
{ "id": "nvidia/parakeet-tdt-0.6b-v3", "object": "model", "created": 0, "owned_by": "nvidia" }
],
"providers": [
{
"id": "parakeet",
"name": "NVIDIA Parakeet",
"models": [{ "id": "parakeet-v3", "name": "Parakeet v3", "description": "...", "available": true }]
}
],
"capabilities": {
"llm": true,
"text_process": true,
"text_suggest": true
}
}
capabilities object:
| Field | Type | Meaning |
|---|---|---|
llm |
bool | LLM_BASE_URL + LLM_MODEL are set and LLM is active |
text_process |
bool | /v1/text/process is open (llm=true AND TEXT_ROUTES_OPEN=true, or auth=true) |
text_suggest |
bool | /v1/text/suggest is open (same condition as text_process) |
text_summarize |
bool | /v1/text/summarize is open (same condition as text_process) |
formatting |
bool | the formatting context key is honoured on cleanup (llm=true) |
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.
- 3d ago First seen · 325 lines · 3,074 tokens per session scan A ad03e436d299
Diction AGENTS.md is an instructions file published in the GitHub repository DictionLabs/Diction (192 stars, last pushed 6d ago), licensed MIT. It adds 3,074 tokens to every session, about $0.0154 per session on Opus 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 2026-08-30.
Other instructions, from other repositories
noisy-coding CLAUDE.md
Instructions for noisy/noisy-coding, covering noisy-coding — agent notes, local development setup, key docs, releasing and restarting the daemon.
Bark CLAUDE.md
Instructions for Finb/Bark, covering claude.md, project overview, development commands, dependencies and install cocoapods dependencies.
maui-labs AGENTS.md
Instructions for dotnet/maui-labs, covering agent instructions, repository overview, products, technology stack and building.
AppClaw CLAUDE.md
Instructions for appclawhq/AppClaw, covering claude.md, what is appclaw?, build & run commands, architecture and entry point & cli modes (src/index.ts).
swift-architecture-skill AGENTS.md
Instructions for efremidze/swift-architecture-skill, covering agents guidelines for swift architecture skill, working with this repository, 1. understanding the skill, 2. modifying architecture playbooks and 3. testing and validation.
brownbear CLAUDE.md
Instructions for DudeAint/brownbear, covering claude.md — engineering operating manual for ai agents, ⚠️ read this before you touch a single line, 0. prime directives (non-negotiable), 1. project identity and what we are building (the 5 modules).