Salvage: uncommitted local work from Macs-MacBook-Pro (1 files)

agents.md

@@ -1,46 +1,8 @@
 # Agent instructions
 
-## Scope and hierarchy
+## Repository-local Git worktrees
 
-- These instructions apply to the whole `zed-pkg/zed-interfaces` repository unless a deeper lowercase `agents.md` adds narrower rules.
-- Before editing, resolve the current working directory and load every readable ancestor `agents.md` from the filesystem root to the working directory. Do not search siblings. Resolve symlinks, deduplicate resolved files, and report unreadable or cyclic instruction files.
-- `.claude/CLAUDE.md`, `.gemini/GEMINI.md`, and `.openai/AGENTS.md` are pointers only. Never duplicate instructions in tool-specific files.
-
-## Repository role
-
-This crate is the shared contract boundary for Zed manifests, lockfiles, registry models, language targets, paths, version requirements, synchronization, and VCS metadata. Changes here affect the CLI, API server, web server, sync engine, and generated clients.
-
-## Working rules
-
-- Preserve serialization compatibility and deterministic ordering unless an explicitly versioned migration says otherwise.
-- Treat public Rust types, JSON schemas, TOML fields, lockfile formats, and path conventions as cross-repository APIs.
-- Add round-trip and negative tests for every parser or schema change; do not weaken validation to make one consumer pass.
-- Update interface consumers and monorepo pins in contract-first order when a change spans repositories.
-- Keep the crate free of service-specific network, credential, database, and deployment policy.
-- Never commit registry tokens, cloud credentials, generated secrets, or production environment files.
-- Keep `Cargo.lock` and `flake.lock` committed; do not allow CI to update either lock implicitly.
-- Pin GitHub Actions by immutable commit SHA and keep workflow permissions read-only unless a documented write is required.
-- Resolve conflicts by preserving serialization compatibility, validation strength, generated-schema determinism, and consumer expectations rather than selecting an entire side.
-
-## Reproducible validation
-
-Use the pinned shell rather than mutable host toolchains:
-
-```sh
-nix develop -c agent-check
-```
-
-Focused stages are available while iterating:
-
-```sh
-nix develop -c agent-check format
-nix develop -c agent-check lint
-nix develop -c agent-check test
-nix develop -c agent-check schemas
-```
-
-The default command runs Nix/workflow preflight, rustfmt, Clippy with warnings denied, unit tests, doctests, schema generation, and a clean-tree schema drift check.
-
-## Validation
-
-The pinned `agents policy` workflow validates this hierarchy and the three tool pointers. Run `nix develop -c agent-check` before requesting review.
+- Create or use a Git worktree only when the human operator explicitly authorizes it for the current task. Concurrency or a dirty checkout is not permission by itself.
+- Put every authorized worktree at `<repository-root>/tmp/worktrees/<name>`; from the repository root, use `./tmp/worktrees/<name>`. Never place worktrees beside repositories or organization directories.
+- Keep `tmp`, `temp`, `tmp/worktrees`, and `temp/worktrees` ignored in the repository-root `.gitignore`. Do not commit files from those directories.
+- Relocate or remove a worktree only when the operator explicitly requests it. Before removal, preserve and publish intended changes, verify its commit is represented on the target branch, and confirm there are no tracked, untracked, ignored-sensitive, or in-use files that must survive. Remove it with `git worktree remove <path>` without `--force`; never delete a worktree directory with `rm`.