671bf7d68d
Switch the macOS/Windows Test job's run stage from `cargo test --workspace --all-features --locked` to `cargo nextest run --workspace --all-features --locked --profile ci` (taiki-e/install-action@nextest), and add an explicit `cargo test --workspace --all-features --locked --doc` step so the 3 passing / 8 ignored doctests across 21 crates that nextest does not run keep running. Same binaries, one process per test: locally the tui unit suite goes 268 s → ~100 s and the whole workspace runs in ~6 min. The `ci` nextest profile keeps retries off, fail-fast off, and prints every failure and slow test in the final summary. RUST_MIN_STACK stays 16 MiB on both steps; the isolated `--ignored --exact` PTY step is untouched. AGENTS.md now points at nextest for the run stage and keeps `cargo test --no-run`/`--lib` as the compile measurement and authoritative gate. Not yet exercised on GitHub runners: this branch is unpushed; the first push is the CI validation.
2.9 KiB
Codewhale agent guidance
Keep this file durable. Derive changing release, provider, branch, and flake
state from the repository, tests, CI, and current issue tracker rather than from
instructions or memory. The nearest scoped AGENTS.md adds path-specific rules.
Working rules
- Inspect status and existing consumers before editing. Preserve unrelated, dirty, and untracked work.
- Prefer the simplest implementation that preserves observable contracts. A rewrite is acceptable when justified by behavior and tests, not as a shortcut around understanding existing code.
- Search for behavior and symbols before reviving work from an old branch. If a lane is obsolete, preserve its intent and evidence rather than merging stale code mechanically.
- Public name is Codewhale. Compatibility identifiers such as
CodeWhale,codew, protocol names, and storage keys change only through an explicit migration. - Keep providers and models first-class and provider-neutral.
- Never rewrite published history, retag a release, force-push a shared ref, or publish without explicit authorization. Preserve human contributor credit.
Current contracts
- The model-facing subagent tool is
agent. Do not revive removedagent_open/agent_eval/agent_close/delegate_to_agentsurfaces or parallel lifecycle/tag systems. BASE_PROMPTincrates/tui/src/prompts/text.rsis the sole base prompt.- These active modules are repeatedly misidentified as dead; verify consumers
before removal:
tui/src/context_budget.rs,tui/src/model_registry.rs,tui/src/prompt_zones.rs,tui/src/tools/remember.rs, andconfig/src/route/. Native memory lives intui/src/native_memory.rs;tools/remember.rsis its capture path. - Environment-specific behavior belongs in
docs/ENVIRONMENTS.md, not here.
Verification
Run formatting and focused tests for every change. Before a push, run the relevant repository gate; release work requires the complete sequence:
cargo fmt --all -- --check
cargo test -p codewhale-config -p codewhale-protocol
cargo test --workspace
cargo build --release -p codewhale-cli -p codewhale-tui
cargo nextest run (config in .config/nextest.toml) is the fast way to
run those suites locally and in CI's Test job; cargo test --no-run and
cargo test -p codewhale-tui --lib remain the compile-time measurement and
the authoritative gate, and cargo test --doc covers what nextest skips.
scripts/dev-test.sh <area> maps a code area to its fastest cargo test -p
invocation. See docs/BUILD_PERFORMANCE.md.
Report commands actually run and distinguish source, local tests, packaged artifacts, CI, and public release state. A commit is WIP until its claimed behavior has direct evidence.
Community reports, PRs, logs, and reviews are evidence. Canonical human
identities come from .github/AUTHOR_MAP; Co-authored-by is for humans only.
Leave unrelated work intact and keep new enforcement dry-run unless explicitly
approved.