mirror of
https://github.com/nearai/ironclaw.git
synced 2026-09-03 08:06:01 +08:00
* docs(guidance): repo-wide agent-guidance audit — fix drift, prune 21.5k lines, consolidate tests/ onto AGENTS.md convention
Full-layer audit of the agent-guidance system (root contracts, .claude rules/
skills/commands, family and crate AGENTS.md, CONTRACT specs, tests guidance,
docs/internal), verified reference-by-reference against HEAD.
- Fix stale/ghost references: UserSandboxProcessPort, ProductSurfaceError,
LlmError::ContextLengthExceeded, INVERTED_PORT_IMPLEMENTORS, split channel
traits (ChannelIngress/ChannelReply/ChannelDelivery), memory-native's
never-implemented EmbeddingProvider seam, wrong layer/crate/module counts.
- Convert unpinned prose numbers to regeneration commands or pinning-test
citations across root, family, and crate guidance (drift-proofing).
- tests/: rename CLAUDE.md -> AGENTS.md with CLAUDE.md symlinks (crates/
convention), extend scripts/ci/check-guidance.py discovery to tests/,
delete stale e2e scenario tables, dedupe tier taxonomy against
.claude/rules/testing.md.
- Commands/skills: delete six dead v1 commands (add-tool, review-pr,
review-crate, fix-issue, respond-pr, add-sse-event) and the v1-teaching
architecture-video skill; convert ironclaw-reborn-skill-maintainer into
the auto-loading rule .claude/rules/guidance-maintenance.md; fix clippy
-D warnings and portable date in surviving commands; triggers-only
frontmatter; add automations section to reborn-feature.
- Rules: rename gateway-events.md -> events.md; revive
scripts/check-type-duplicates.py (glob matched zero types since the
family reorg); index all 15 rules in root AGENTS.md for Codex parity.
- docs/internal: delete 70 superseded plans/specs/design docs (~21.5k
lines, each re-verified unreferenced); fix misleading v1-migration
status lines; rewrite the contracts index as a recipe; restore two docs
that proved live-referenced.
- Trim composition CONTRACT.md route-mirror sections (invariants kept).
Verified: check-guidance.py (384 files, 0 grandfathered),
docs_publication_boundary.py, cargo test -p ironclaw_architecture_tests,
scripts/ci test-plan suite (87/87) — all green.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* docs(guidance): address PR #7797 review comments
Verified each of the ~40 bot/reviewer findings against the tree; applied the
valid mechanical fixes, rebutted the rest with evidence (see PR comment).
- Test planner: add renamed tests/ guidance aliases to IGNORED_GUIDANCE_PATHS
(reproduced the fail-closed abort on this PR's own changed-file list) and
extend the planner test to all six guidance paths.
- check-guidance: add self-tests proving tests/-tree discovery and alias
enforcement (48 tests, was 46); new self-test file for
check-type-duplicates.py (4 tests).
- Portability: replace GNU-only date -d in triage commands with a python3
one-liner (works on macOS BSD and Linux).
- Count/claim accuracy: product_contracts manager-port prose 4 -> 3 (matches
INVERTED_PORT_IMPLEMENTORS), kernel grep -cF for literal #[test] (was regex
char class, 2 vs 23), rg -o|wc -l for a true total in architecture.md,
Rust-scoped LlmProvider count (catches 5 generic impls), measured
1/73-crate dual-backend claim in pr-shepherd, executable wc -l in
assistant guidance, AST/pytest recipes for the e2e test-count figures.
- Content: deslop co-author line no longer hardcodes an address; ship.md
surfaces Postgres-skip counts; risk-label guidance documents the crates/**
labeler blind spot; e2e authoring recipe leads with reborn_v2_* fixtures;
stale CLAUDE.md line citation replaced with a stable anchor; ✎ provenance
notes for two deleted-plan citations; unified-channel-model status text
reconciled with an explicit ChannelDelivery-only exception note.
Gates: check-guidance (384 files, 0 grandfathered), docs boundary, planner
tests 87/87, check-guidance self-test 48/48, type-dup self-test 4/4 - green.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* fix(ci): classify scripts/test-check-type-duplicates.py in the PR test planner
The self-test added in 77bc8f8c5 was never registered in
PR_STATIC_CONTROL_PATHS, so the planner's fail-closed unmapped-path arm
aborted 'Detect Reborn test scope' and cascaded into the whole Reborn
matrix skipping. Classified like its subject (deliberately CI-unwired
local dev tool, per the existing entry's rationale) and pinned in the
static-control planner test alongside it.
Verified: planner tests OK; planner run against this PR's full
changed-file list now returns mode=selected with the path owned by
static checks.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* fix(docs): drop parity-QA path reference from tests/integration/AGENTS.md
The guidance dedup in this PR added a literal
tests/support/reborn_parity_qa reference to tests/integration guidance,
which scripts/ci/check-test-suite-boundaries.sh correctly flags: the
one-way dependency guard covers docs too, and origin/main's version of
this file carried no such reference. Fix the content, not the check -
the tier comparison is reworded to describe the RebornBinaryE2EHarness
seam difference without naming the parity/QA tree.
Verified: check-test-suite-boundaries.sh OK; check-guidance OK; the
full 'Detect Reborn test scope' job reproduced locally end-to-end.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* ci: run the type-duplicates self-test in Code Style; exercise the production report path
Address the two follow-up review findings on 77bc8f8c5/9ed30caf8:
- Wire scripts/test-check-type-duplicates.py into code_style.yml's
Static-check self-tests step (next to test-check-guidance.py) so the
regression test is actually enforced; update the planner's
PR_STATIC_CONTROL_PATHS comment accordingly (classification unchanged
- Code Style is the static lane).
- Strengthen the semantic-duplicate self-test to also drive the
production main() report path and assert on its printed candidate
output, instead of only re-computing similarity locally. Strictly
stronger; the other three tests are untouched.
Verified: self-test 4/4, planner tests 87/87, workflow-contract
self-test 94/94, planner simulation over both changed paths classifies
cleanly (no unmapped-path abort).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* ci: trigger fast-checks on type-duplicates script changes; symbol recipe for the tier reference
- code_style.yml's has_code filter covered scripts/ci/ but not the bare
scripts/ type-duplicates pair, so a diff touching only the new
self-test skipped the fast-checks job that runs it (the previous
commit's 'runs unconditionally' claim was wrong - corrected the
planner comment too). Added the two files to the filter following the
check_no_panics precedent and pinned them as in_scope probes in the
ws12 workflow-contracts routing test.
- tests/integration/AGENTS.md tier reference is now re-verifiable via a
single-hit symbol recipe (rg 'struct RebornBinaryE2EHarness') instead
of a path citation, which the test-suite boundary guard forbids from
this subtree.
Verified: ws12 workflow contracts 94/94, planner tests 87/87,
check-test-suite-boundaries OK, check-guidance OK.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
---------
Co-authored-by: Henry Park <16583448+henrypark133@users.noreply.github.com>
Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
173 lines
7.4 KiB
Markdown
173 lines
7.4 KiB
Markdown
# Contributing
|
|
|
|
## Getting Started
|
|
|
|
```bash
|
|
git clone https://github.com/nearai/ironclaw.git
|
|
cd ironclaw
|
|
./scripts/dev-setup.sh
|
|
```
|
|
|
|
This installs the Rust toolchain, WASM targets, git hooks, and runs initial checks.
|
|
|
|
## How to Contribute
|
|
|
|
- Bug fixes, docs improvements, and focused cleanup tied to a concrete problem are welcome.
|
|
- Search existing issues and PRs before opening a new one to avoid duplicates.
|
|
- Keep changes scoped. One bug, one feature, or one documentation improvement per PR.
|
|
|
|
### Creating Issues
|
|
|
|
Open an issue when you are reporting a bug, proposing a feature, or documenting a gap in behavior.
|
|
|
|
For bug reports, include:
|
|
|
|
- What you expected to happen
|
|
- What actually happened
|
|
- Clear reproduction steps
|
|
- Relevant logs, screenshots, or error output
|
|
- Environment details when they matter (OS, database backend, feature flags, commit/branch)
|
|
|
|
For feature requests:
|
|
|
|
- Open an issue first before writing code
|
|
- Explain the problem being solved, not just the implementation idea
|
|
- Wait for maintainer feedback before investing in a large PR
|
|
|
|
We require an issue for new features so maintainers can prioritize the work and confirm it fits the roadmap before anyone spends time implementing it.
|
|
|
|
### Fixing Bugs
|
|
|
|
- Small, targeted bug-fix PRs are welcome
|
|
- If there is already an issue, link it in your PR
|
|
- If the bug is non-trivial, security-sensitive, or changes behavior across subsystems, open or confirm an issue first so the approach can be aligned before implementation
|
|
|
|
### Refactor-Only PRs
|
|
|
|
Refactor-only PRs are not accepted from contributors outside the core team. If a refactor is necessary to land a bug fix or approved feature, keep it minimal and clearly tied to that change.
|
|
|
|
## Development Workflow
|
|
|
|
Use the [Ironclaw TDD playbook](docs/internal/testing-playbook.md) to choose the
|
|
smallest test layers that prove a feature or bug fix and to find the correct
|
|
test location.
|
|
|
|
```bash
|
|
cargo fmt # format
|
|
cargo clippy --all --benches --tests --examples --all-features # lint (warnings surface here; the pre-PR block below denies them)
|
|
cargo test # unit + integration suites (Postgres legs self-provision testcontainers; skipped when Docker is unavailable)
|
|
```
|
|
|
|
These commands are for day-to-day iteration while you are developing locally. The pre-submission checks below are intentionally stricter and use CI-style flags so you can catch formatting drift and clippy warnings before requesting review.
|
|
|
|
## Before You Open a PR
|
|
|
|
Run the local validation checks required before requesting a review. These are stricter than the commands for iterative development:
|
|
|
|
```bash
|
|
cargo fmt --all -- --check
|
|
cargo clippy --all --benches --tests --examples --all-features -- -D warnings
|
|
cargo build
|
|
cargo test
|
|
```
|
|
|
|
Also run the owning crate's feature-gated suite when your change touches
|
|
database-backed or runtime-integration behavior (the workspace-root
|
|
`integration` feature is empty — the flag only means something per-crate):
|
|
|
|
```bash
|
|
cargo test -p <owning-crate> --features integration # e.g. -p ironclaw_hooks for the Postgres/libSQL parity matrix
|
|
```
|
|
|
|
Before asking for review:
|
|
|
|
- Build and exercise the changed path locally, not just the narrowest unit test
|
|
- Keep the PR focused and avoid mixing unrelated concerns
|
|
- Fill out the PR template with a clear summary, validation notes, and impact assessment
|
|
- If your change affects tracked behavior, update `FEATURE_PARITY.md` in the same branch
|
|
- If onboarding or setup behavior changes, update the relevant setup docs in the same branch
|
|
- If you are using a coding agent and it supports them, run `pr-shepherd` before opening or updating the PR
|
|
- For bug-issue fixes there is no dedicated repo command: follow the test-first discipline in `AGENTS.md` (pin the bug in a failing test at the right tier, then fix), and let CI's review bot plus `codex review` cover the review pass
|
|
- `codex review --base origin/main` is also encouraged before requesting review
|
|
|
|
## Review Follow-Through
|
|
|
|
Review conversations are author-owned.
|
|
|
|
- Address each review comment with a code change or a clear explanation
|
|
- Resolve conversations you have handled; leave them open only when reviewer judgment is still needed
|
|
- Do not leave review cleanup for maintainers when the follow-through belongs to the author
|
|
|
|
If a PR is stale for more than 48 hours after review feedback is posted, maintainers may take over the follow-up work and land the changes needed to accomplish the original PR or issue intent.
|
|
|
|
## Code Style
|
|
|
|
- Zero clippy warnings policy
|
|
- No `.unwrap()` or `.expect()` in production code (tests are fine)
|
|
- Use `thiserror` for error types, map errors with context
|
|
- Prefer `crate::` for cross-module imports
|
|
- Comments for non-obvious logic only
|
|
|
|
See `CLAUDE.md` for full style guidelines.
|
|
|
|
## Feature Parity Requirement
|
|
|
|
When your change affects a tracked capability, update `FEATURE_PARITY.md` in the same branch.
|
|
|
|
### Required before opening a PR
|
|
|
|
1. Review the relevant parity rows in `FEATURE_PARITY.md`.
|
|
2. Update status/notes if behavior changed.
|
|
3. Include the `FEATURE_PARITY.md` diff in your commit when applicable.
|
|
|
|
## Review Tracks
|
|
|
|
All PRs follow a risk-based review process:
|
|
|
|
| Track | Scope | Requirements |
|
|
|-------|-------|-------------|
|
|
| **A** | Docs, tests, chore, dependency bumps | 1 approval + CI green |
|
|
| **B** | Features, maintainer-requested refactors, new tools/channels | 1 approval + CI green + test evidence |
|
|
| **C** | Security (`crates/substrates/ironclaw_safety/`, `crates/substrates/ironclaw_secrets/`), turn runtime and agent loop (`crates/kernel/`, `crates/loop/`), database schema, CI workflows | 2 approvals + rollback plan documented |
|
|
|
|
Select the appropriate track in the PR template based on what your changes touch.
|
|
|
|
## Database Changes
|
|
|
|
IronClaw uses dual-backend persistence (PostgreSQL + libSQL) behind the `RootFilesystem` mount catalog. All new persistence features must support both backends. See `crates/substrates/ironclaw_filesystem/CONTRACT.md` and `.claude/rules/database.md`.
|
|
|
|
## Adding Dependencies
|
|
|
|
Run `cargo deny check` before adding new dependencies to verify license compatibility and check for known advisories.
|
|
|
|
## Document your Changes
|
|
|
|
- The folder `/docs` contains user-facing documentation for technical savvy users, developers and operators. It is built with Mintlify and rendered on the website.
|
|
- The live site deploys from the `docs-live` branch, which release automation points at each stable release — docs merged to `main` appear on the site with the next stable release, not immediately. To fix a wrong page on the live site mid-week, see the docs-hotfix recipe in `docs/internal/weekly-release-strategy.md` ("Docs publication").
|
|
- For features, update the relevant capability doc in `docs/capabilities/`
|
|
- For channels, update the relevant channel doc in `docs/channels/`
|
|
- For extensions / tools, update the relevant doc in `docs/extensions/`
|
|
- Core features live in `docs/capabilities`
|
|
|
|
In case you want to document the library itself (i.e. reference documentation) for other core contributors, use the `docs/internal/` folder
|
|
|
|
If you use your Claude Code to "plan" and want to leave a record of it, use the `docs/internal/plans` folder.
|
|
|
|
### Skills
|
|
Read the `.claude/skills/mintlify-docs` for guidelines on how to generate documentation with mintlify.
|
|
|
|
### Test the Docs
|
|
To make sure the documentation still works, do:
|
|
|
|
```bash
|
|
cd docs
|
|
mint dev
|
|
```
|
|
|
|
To make sure you did not break any internal links, do:
|
|
|
|
```bash
|
|
cd docs
|
|
mint broken-links
|
|
```
|