lshed 0.12.0 → 0.13.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
package/CHANGELOG.md CHANGED
@@ -1,5 +1,19 @@
1
1
  # Changelog
2
2
 
3
+ ## 0.13.0 — 2026-09-05
4
+
5
+ `--agent agy` places into Google Antigravity (the `agy` CLI and the IDE).
6
+
7
+ - Skills go to `~/.gemini/config/skills/`, the one global location both the Antigravity IDE and CLI read. Instructions are concatenated into `~/.gemini/AGENTS.md`, one level above that root: agy reads both `AGENTS.md` and `GEMINI.md` there, and Gemini CLI also owns `GEMINI.md`, so lshed leaves that one alone. MCP servers go to `~/.gemini/config/mcp_config.json` in agy's shape (no `type`, `serverUrl` for http); `${VAR}` is filled by `restore`.
8
+ - A part above the root (`../AGENTS.md`) is backed up under `lshed/backups/<stamp>/__/` so backups never escape the backup directory.
9
+ - Verified on this machine with agy 1.1.26 through the VM probe (`scripts/vm/probe.sh agy`): skill, rules file, MCP list, and a linked skill all read. agy does not read `~/.agents/skills`, so it is not part of the `agents` target.
10
+
11
+ ## 0.12.1 — 2026-09-05
12
+
13
+ - Fix: with `CLAUDE_CONFIG_DIR` set on a machine where Claude Code has not run yet, `restore` wrote MCP servers to `<dir>.json` next to the config dir instead of `<dir>/.claude.json` inside it, where Claude Code actually reads them. Found by the VM probe below. Without `CLAUDE_CONFIG_DIR` nothing changes.
14
+
15
+ VM probe for the other agents (`scripts/vm/`): `install-tools.sh` bakes an image with Codex, Gemini CLI, Copilot CLI, Cursor CLI and lshed; `cloud-init.yaml` injects API keys at boot and runs `probe.sh`, which restores a throwaway shed into each tool's real root and asks the tool, non-interactively, for a passphrase kept in a skill, a codeword kept in the instructions file, and the same skill again after `--link`. Where a tool can parse its own config without a model (`codex mcp list`, `gemini mcp list`, `codex debug prompt-input`) that is checked too. Run on this machine against Codex 0.153.2: Codex reads `$CODEX_HOME/skills` (deprecated in its source; the docs now name only `~/.agents/skills`), `$CODEX_HOME/AGENTS.md`, the `[mcp_servers.*]` table lshed writes, and a linked skill.
16
+
3
17
  ## 0.12.0 — 2026-09-04
4
18
 
5
19
  MCP servers follow the shed into the other agents.
package/README.md CHANGED
@@ -247,12 +247,13 @@ Instructions fragments are ordered. `restore` writes a `CLAUDE.md` that `@`-impo
247
247
 
248
248
  ### Other agents, same shed
249
249
 
250
- Codex, Gemini CLI, Copilot CLI and Cursor all read skills from `<their config dir>/skills/<name>/SKILL.md`, the same layout Claude Code uses, and all of them also read the shared `~/.agents/skills/`. So one shed can serve them all. Pick the target with `--agent`:
250
+ Codex, Gemini CLI, Copilot CLI, Cursor and Google Antigravity (`agy`) all read skills from `<their config dir>/skills/<name>/SKILL.md`, the same layout Claude Code uses, and all but Antigravity also read the shared `~/.agents/skills/`. So one shed can serve them all. Pick the target with `--agent`:
251
251
 
252
252
  ```
253
253
  lshed restore --agent agents # ~/.agents/skills: every tool that follows the convention reads it
254
254
  lshed restore --agent codex # ~/.codex/skills + ~/.codex/AGENTS.md
255
255
  lshed restore --agent gemini --link # ~/.gemini/skills + ~/.gemini/GEMINI.md, as links
256
+ lshed restore --agent agy # ~/.gemini/config/skills + ~/.gemini/AGENTS.md (Antigravity IDE and CLI)
256
257
  ```
257
258
 
258
259
  | `--agent` | root | skills | instructions file | MCP servers |
@@ -262,9 +263,10 @@ lshed restore --agent gemini --link # ~/.gemini/skills + ~/.gemini/GEMINI.md
262
263
  | `gemini` | `~/.gemini` | yes | `GEMINI.md`, concatenated | `settings.json` |
263
264
  | `copilot` | `~/.copilot` or `$COPILOT_HOME` | yes | `copilot-instructions.md`, concatenated | `mcp-config.json` |
264
265
  | `cursor` | `~/.cursor` | yes | none (Cursor's user rules live in its settings UI) | `mcp.json` |
266
+ | `agy` | `~/.gemini/config` | yes | `../AGENTS.md`, concatenated (agy reads `GEMINI.md` too, but Gemini CLI owns that one) | `mcp_config.json` |
265
267
  | `agents` | `~/.agents` | yes | none | none |
266
268
 
267
- MCP entries are stored in the shed in Claude Code's shape and translated on the way out: Gemini gets `httpUrl` and no `type`, Copilot gets `type: local` and `tools: ["*"]`, Cursor gets `${env:VAR}` placeholders and Codex gets `env_vars` / `bearer_token_env_var` / `env_http_headers` with the variable *names*, so for those two the secret values never touch the config file. Gemini and Copilot do not expand placeholders, so lshed fills them from your shell at `restore`. Codex's `config.toml` is edited table by table; your comments and other settings stay as they are. `lshed init --agent gemini` reads the same files back into the shed shape, secrets masked.
269
+ MCP entries are stored in the shed in Claude Code's shape and translated on the way out: Gemini gets `httpUrl` and no `type`, Antigravity gets `serverUrl`, Copilot gets `type: local` and `tools: ["*"]`, Cursor gets `${env:VAR}` placeholders and Codex gets `env_vars` / `bearer_token_env_var` / `env_http_headers` with the variable *names*, so for those two the secret values never touch the config file. Gemini, Antigravity and Copilot do not expand placeholders, so lshed fills them from your shell at `restore`. Codex's `config.toml` is edited table by table; your comments and other settings stay as they are. `lshed init --agent gemini` reads the same files back into the shed shape, secrets masked.
268
270
 
269
271
  Each agent root keeps its own `lshed/state.json`, so restoring into `~/.codex` never touches what lshed placed in `~/.claude`, and each can use a different profile or `--link` choice. Parts the target does not understand are announced and skipped: a profile with settings keys restores into Codex without them, with a line saying `codex 은 settings 를 다루지 않아 건너뜁니다`. Claude plugin packages are skipped the same way; `github:`/`git:` packages are cloned into every agent root that restores the profile, so give the other agents a profile without them if that is not what you want. `lshed init --agent codex` works too, and a shed made from Codex restores into Claude Code with `CLAUDE.md` generated from the same fragments. The shed's `agent:` is only a default for `--agent` (`$LSHED_AGENT` also works).
270
272