agentmods.dev

OpenWhisperer AGENTS.md

OpenWhisperer AGENTS.md is an instructions file for Codex, OpenCode from PerIPan/OpenWhisperer. It costs 8,497 tokens per session, scanned D, original, MIT.

Repository instructions for OpenWhisperer, a Swift application with a graphical app and a text-to-speech server. Text-to-speech means turning written words into spoken audio.

In plain words
What is it for?
They guide Swift builds, release bundle creation, unit and integration tests, the headless speech server, and command-line speech checks.
Why use it?
They tell coding agents how to build, test, package, and manually check the project, including its macOS app and speech server.

Instructions file for CodexOpenCode

not rated 5 changed today D scan 8,497 tokens original MIT
View the source file PerIPan/OpenWhisperer
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 instructions/peripan/openwhisperer/agents-md
Clone the repo
git clone --depth 1 https://github.com/PerIPan/OpenWhisperer

Made for: Codex, OpenCode.

Wrote 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.

agentmods badge for OpenWhisperer AGENTS.md

README.md
[![agentmods](https://agentmods.dev/badge/instructions/peripan/openwhisperer/agents-md.svg)](https://agentmods.dev/instructions/peripan/openwhisperer/agents-md)
Your own site
<a href="https://agentmods.dev/instructions/peripan/openwhisperer/agents-md"><img src="https://agentmods.dev/badge/instructions/peripan/openwhisperer/agents-md.svg" alt="Measured on agentmods" height="20"></a>
Per session 8,497 This file is loaded in full into every session.
When invoked 8,497 The same file — it is already loaded in full.
Security scan D 3 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 $0.08497 $0.08497
Opus 5 $0.04248 $0.04248
Sonnet 5 $0.01699 $0.01699
Haiku 4.5 $0.00850 $0.00850

Measured today against content hash b43857da74e8, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade D, and why

OpenWhisperer AGENTS.md scanned grade D with 3 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 today.

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.

Reads agent configuration directoriesmediumAgent snooping

.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.

**One hook, both platforms.** `voice-context.sh` serves Claude Code *and* Codex — both fire a `UserPromptSubmit` command hook whose stdin carries `prompt`+`session_id` and whose stdout `{hookSpecificOutput:{additionalCon

Recursive force deletehighDestructive command

rm -rf with a variable or a broad path is one typo away from removing the wrong tree.

- **Rebuilding & Installing locally:** Since hook files (`hooks/*.sh`) are bundled into the application resource path (`OpenWhisperer.app/Contents/Resources/hooks/`) during build, any hook changes require rebuilding the

Makes network callslowCapability

Not a fault in itself. Listed so you know the mod talks to something, and to what.

swift run HookTests # bash-hook integration tests (stubbed curl + temp HOME)
AGENTS.md · 152 lines

How it starts

The opening of the file, as written. The whole thing — 152 lines — stays where its author put it; the contents beside it link to each section on GitHub.

AGENTS.md

This file defines the project-scoped rules, workflows, commands, and architecture details for AI agents working in the OpenWhisperer repository.

Commands

All Swift commands run from the app/ directory (it's the SwiftPM package root).

cd app
swift build                                           # debug build
swift build -c release                                # release build
OW_SIGN_IDENTITY="OpenWhisperer Dev" ./build-dmg.sh   # build signed bundle + .dmg for local dev (keeps TCC grants)
./build-dmg.sh                                        # build ad-hoc bundle + .dmg (release) into app/.build/

swift run OpenWhispererKitTests      # pure-logic unit tests (exits non-zero on failure)
swift run HookTests                  # bash-hook integration tests (stubbed curl + temp HOME)

swift run OpenWhisperer --serve-tts  # headless TTS server on :8000 (no GUI) for curl/CI; TTS_PORT overrides

Manual TTS smoke checks (server must be running, GUI or --serve-tts):

curl http://localhost:8000/v1/models
echo "hello" | scripts/speak.sh

Testing notes

  • This machine has Command Line Tools only, no full Xcode, so there is no XCTest/swift-testing. Both test targets are plain @main executables that aggregate *Failures() -> [String] check groups and exit(1) if any fail. There is no per-test filter — to narrow a run, comment out check-group calls in the runner (Tests/OpenWhispererKitTests/SubmitTriggerTests.swift or Tests/HookTests/main.swift).
  • Only pure logic (no AppKit/AVFoundation/WhisperKit/FluidAudio) is unit-testable, and it lives in the OpenWhispererKit target precisely so it builds fast under CLT. Put new testable logic there.

Working on changes

The loop this repo already follows. The parent checkout stays on main; feature branches live only under .claude/worktrees/ — never branch in place.

Pick the lightest safe path:

  • Local (commit straight to main, no PR) — docs/CLAUDE.md/comment edits, or a small, self-contained, low-risk code fix where a PR adds ceremony without safety. Stay on main, make the smallest edit, verify (readback for docs; the test targets for code), commit, push. If it unexpectedly grows beyond small/isolated, switch to the PR path before committing.
  • PR — multiple files, real logic, a dependency change, or user-visible behavior:
    1. Worktree off main: git worktree add .claude/worktrees/<slug> -b <slug> (or EnterWorktree in Claude Code).
    2. Build + test from app/: swift run OpenWhispererKitTests and swift run HookTests (both exit(1) on failure). For a packaged .app/.dmg, use OW_SIGN_IDENTITY="OpenWhisperer Dev" ./build-dmg.sh locally so TCC grants survive the rebuild (see Conventions).
    3. PR: rebase onto origin/main if it moved, push, gh pr create.
    4. After merge: sync main (git pull --ff-only), git worktree remove .claude/worktrees/<slug>, delete the branch. Leave no orphaned worktree dir — a stale bundle holding :8000 is a known foot-gun.

Read the full file on GitHub · 152 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. today Changed · +2 lines · +818 tokens per session b43857da74e8
  2. 4d ago First seen · 150 lines · 7,679 tokens per session scan D 382fbc3843b3

Subscribe to this mod's changes

OpenWhisperer AGENTS.md is an instructions file published in the GitHub repository PerIPan/OpenWhisperer (5 stars, last pushed yesterday), licensed MIT. It adds 8,497 tokens to every session, about $0.0425 per session on Opus 5. A static security scan graded it D with 3 findings (reads agent configuration directories, recursive force delete, makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Apple SiliconGit and workflowProductivityTestingaihands-free-voice-interfacemlxopen-sourcestt-tts
Related

Other instructions, from other repositories

speech-swift AGENTS.md

AGENTS.md instructions for soniqo/speech-swift, covering agent instructions, communication style, workflow, running tests and benchmarks — sequential and memory-aware and git conventions.

soniqo/speech-swift · 5,053 tokens

Vocello AGENTS.md

AGENTS.md instructions for PowerBeef/Vocello, covering agents.md — vocello (qwenvoice), product and authority, start and resume work, hard invariants and domain routing.

PowerBeef/Vocello · 5,217 tokens

Logue CLAUDE.md

Instructions for bitwize-ai/Logue, covering logue — development guidelines, project overview, running the build you just made, where documentation goes and dependencies.

bitwize-ai/Logue · 7,535 tokens

speech-studio AGENTS.md

Instructions for soniqo/speech-studio, covering agent instructions, project, stack, sibling repos under /repos/ and build.

soniqo/speech-studio · 4,490 tokens

localvoxtral AGENTS.md

AGENTS.md instructions for T0mSIlver/localvoxtral, covering localvoxtral — agent guide, build & test — read this first on a non-mac dev box, proof culture — non-negotiable, test tiers — the short version and ci / shipping.

T0mSIlver/localvoxtral · 2,917 tokens

OpenSuperMLX AGENTS.md

AGENTS.md instructions for axot/OpenSuperMLX, covering agents.md — opensupermlx, agent behavioral guidelines, 1. think before coding, 2. simplicity first and 3. surgical changes.

axot/OpenSuperMLX · 5,990 tokens