Observed since 20 Oct, 22:35 UTC
No superseded metadata states collected. Unchanged observations do not create new revisions.
Reliable coding-agent commands across Unicode paths, shell encodings, Git output, Python launchers, and mixed-encoding files.
agent-locale helps coding agents run project commands reliably across Unicode paths, shell encodings, Git output, Python launchers, and mixed-encoding files.
It can:
A coding-agent CLI is optional. The core commands work without Codex, Claude Code, OpenCode, or Gemini CLI.
git clone https://github.com/starsaround/agent-locale.git
cd agent-locale
python -m venv .venv
Activate the environment:
# macOS / Linux
source .venv/bin/activate
# Windows PowerShell
.\.venv\Scripts\Activate.ps1
Install the CLI:
python -m pip install .
agent-locale --help
Install directly from GitHub without cloning:
python -m pip install "agent-locale @ git+https://github.com/starsaround/agent-locale.git"
On Windows, use py -3 or an absolute interpreter path if python opens the Microsoft Store. On macOS or Linux, use a virtual environment when the system Python rejects package installation.
Inspect the current machine and repository:
agent-locale doctor
agent-locale quickstart
Create the repository-local command gateway:
agent-locale bootstrap
agent-locale repair verify
agent-locale exec -- python --version
List and run saved project commands:
agent-locale commands list
agent-locale run test
Use agent-locale doctor --help for command-specific options, replacing doctor with the command you want to inspect.
Generate the optional harness pack when you need reusable pack files in the current repository:
agent-locale export-pack
bootstrap may create:
.agent-locale/ for gateway state, reports, and command logs;.agents/skills/i18n-reliability/ for the managed repository skill;AGENTS.md.It does not change global PATH, shell profiles, or system locale settings.
Preview cleanup:
agent-locale clean --all --installed-skill --dry-run
agent-locale deactivate --update-agents --remove-logs --remove-installed-skill --dry-run
Remove --dry-run after reviewing the planned changes.
Core diagnosis, gateway, repair, cleanup, and edit verification work without an agent CLI.
| Capability | Supported agents |
|---|---|
| CLI discovery | Codex, Claude Code, OpenCode, Gemini CLI |
| Task templates and event summaries | Codex, Claude Code, OpenCode |
| Optional edit-preservation self-test | Codex |
Agent-specific adapters are optional. Select one only when generating a task template or running an agent-specific self-test.
| Platform | Status |
|---|---|
| Windows 11 | Tested |
| macOS on Apple silicon | Tested |
| Linux | Limited |
| Windows 10 | Unverified |
| macOS on Intel | Unverified |
Commands passed to agent-locale exec run with the current user's permissions. Review commands before running them.
Full command logging may contain sensitive output even though common secret formats are redacted. Use one of these modes for sensitive commands:
agent-locale exec --log-policy metadata -- python -V
agent-locale exec --log-policy none -- python -V
Replace python -V with the command you need to run.
Reports, logs, fixtures, and repository state remain local unless you explicitly share or upload them. Review generated files before attaching them to a public issue.
Use GitHub Issues for bugs and compatibility feedback. Do not post credentials, private source code, proprietary logs, or other sensitive information.
Download counts and popularity will appear here as the collector receives them.
| Version | Published | Direct dependencies | Open version |
|---|---|---|---|
| 0.1.0a1 | 18 Jul 2026 | 2 |
This describes our collected data; it is not a security assessment.
Observed since 20 Oct, 22:35 UTC
No superseded metadata states collected. Unchanged observations do not create new revisions.
package:ef5cf42439492287607a7d2cce72fed70ebc3c7ff29cbf4f9a26e310f2a2b96b