Skip to content

[codex] tighten steward strategy#66

Merged
lbliii merged 2 commits intomainfrom
codex/tighten-steward-strategy
May 3, 2026
Merged

[codex] tighten steward strategy#66
lbliii merged 2 commits intomainfrom
codex/tighten-steward-strategy

Conversation

@lbliii
Copy link
Copy Markdown
Owner

@lbliii lbliii commented May 3, 2026

Summary

Tightens the steward system so future bugbashes and cross-boundary changes produce contract-oriented findings, explicit synthesis, and reliable collateral updates.

What changed

  • Added root steward guidance for contract checklists, evidence-backed finding format, synthesis requirements, collateral rules, and parity-matrix expectations.
  • Added ## Contract Checklist sections to every scoped AGENTS.md steward file.
  • Made each steward spell out the surfaces that should move with code: tests, docs, examples, scaffold, benchmarks, changelog, and generated artifacts where applicable.
  • Added a changelog fragment for the process change.

Why

The project-wide bugbash showed the steward model is useful when it turns a local bug into a full contract repair across code, tests, docs, and onboarding surfaces. This PR makes that behavior repeatable instead of relying on reviewer instinct.

Validation

  • git diff --check passed
  • Verified all AGENTS.md files include ## Contract Checklist
  • make docs-test passed

Steward Notes

Consulted root and all scoped steward files.

Decisions:

  • Kept this as instruction/documentation only; no runtime code changes.
  • Added local checklists rather than a separate process document so agents see the right checklist at edit time.
  • Required no collateral: <reason> style notes when tests/docs/examples do not need to move.

Risks:

  • More process text can be ignored if too vague, so the additions are concrete checklist items and finding fields.

Follow-up:

  • Future swarm PRs should include the parity matrix in their PR body when behavior spans CLI, programmatic call, MCP, schema, docs, examples, or scaffold.

@github-actions
Copy link
Copy Markdown

github-actions Bot commented May 3, 2026

Coverage Report

88.1% overall coverage

File Coverage
src/milo/\_\_init\_\_.py 100.0%
src/milo/\_cells.py 85.9%
src/milo/\_child.py 91.8%
src/milo/\_cli\_help.py 90.5%
src/milo/\_command\_defs.py 99.2%
src/milo/\_compat.py 63.3%
src/milo/\_errors.py 92.8%
src/milo/\_jsonrpc.py 81.4%
src/milo/\_mcp\_router.py 100.0%
src/milo/\_protocols.py 100.0%
src/milo/\_scaffold/\_\_init\_\_.py 100.0%
src/milo/\_types.py 100.0%
src/milo/app.py 58.9%
src/milo/cli.py 87.3%
src/milo/commands.py 89.3%
src/milo/completions.py 96.0%
src/milo/components\_cli.py 88.1%
src/milo/config.py 87.3%
src/milo/context.py 86.6%
src/milo/dev.py 91.9%
src/milo/doctor.py 89.9%
src/milo/flow.py 96.5%
src/milo/form.py 91.7%
src/milo/gateway.py 68.9%
src/milo/groups.py 93.5%
src/milo/help.py 100.0%
src/milo/input/\_\_init\_\_.py 100.0%
src/milo/input/\_platform.py 77.8%
src/milo/input/\_reader.py 96.2%
src/milo/input/\_sequences.py 100.0%
src/milo/live.py 0.0%
src/milo/llms.py 78.8%
src/milo/mcp.py 88.9%
src/milo/middleware.py 100.0%
src/milo/observability.py 100.0%
src/milo/output.py 90.5%
src/milo/pipeline.py 90.1%
src/milo/plugins.py 100.0%
src/milo/reducers.py 100.0%
src/milo/registry.py 89.1%
src/milo/schema.py 92.7%
src/milo/state.py 91.7%
src/milo/streaming.py 100.0%
src/milo/templates/\_\_init\_\_.py 100.0%
src/milo/testing/\_\_init\_\_.py 100.0%
src/milo/testing/\_mcp.py 100.0%
src/milo/testing/\_record.py 86.7%
src/milo/testing/\_replay.py 87.1%
src/milo/testing/\_snapshot.py 100.0%
src/milo/theme.py 100.0%
src/milo/verify.py 84.8%
src/milo/version\_check.py 60.6%

@lbliii lbliii marked this pull request as ready for review May 3, 2026 18:55
Copilot AI review requested due to automatic review settings May 3, 2026 18:55
@lbliii lbliii merged commit 5c46790 into main May 3, 2026
7 checks passed
Copy link
Copy Markdown

Copilot AI left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pull request overview

This PR tightens the repository’s steward/agent guidance by adding explicit “Contract Checklist” sections across the root and scoped AGENTS.md files, aiming to make cross-surface contract repairs (tests/docs/examples/scaffold/etc.) more consistent and evidence-driven.

Changes:

  • Added root-level steward guidance for contract checklists, parity matrices, synthesis requirements, and collateral update expectations.
  • Added ## Contract Checklist sections to scoped steward files (tests, templates, input, scaffold, core, docs, examples, site, benchmarks).
  • Added a towncrier changelog fragment documenting the process change.

Reviewed changes

Copilot reviewed 11 out of 11 changed files in this pull request and generated 3 comments.

Show a summary per file
File Description
AGENTS.md Adds root contract checklist + steward signal/synthesis requirements and done criteria updates.
tests/AGENTS.md Adds test-domain contract checklist (regressions, parity matrices, free-threading coverage notes).
src/milo/templates/AGENTS.md Adds template-domain contract checklist (template checks, strict-undefined, rendering evidence).
src/milo/input/AGENTS.md Adds input-domain contract checklist (decoder fixtures, raw-mode cleanup, non-TTY behavior).
src/milo/_scaffold/AGENTS.md Adds scaffold-domain contract checklist (generated path testing and collateral alignment).
src/milo/AGENTS.md Adds core-domain contract checklist (CLI/programmatic/MCP parity, schema coverage, benchmarks).
site/AGENTS.md Adds site-docs contract checklist (docs parity, snippet tagging, link/navigation validity).
examples/AGENTS.md Adds examples contract checklist (drift checks, template compilation, smoke tests).
docs/AGENTS.md Adds agent-docs contract checklist (quickstart/testing doc updates, snippet drift expectations).
benchmarks/AGENTS.md Adds benchmarks contract checklist (when to add/update benchmarks and evidence expectations).
changelog.d/steward-contract-checklists.changed.md Towncrier fragment describing the process change.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Comment thread AGENTS.md
- User Impact: how humans, agents, or downstream CLIs experience the bug or drift.
- Required Fix: the smallest behavior or docs change that restores the invariant.
- Required Proof: tests, docs checks, snippets, benchmarks, or manual checks that must move with the fix.
- Collateral: docs, examples, scaffold, llms.txt, changelog, migration notes, or benchmarks that also need updates; write "none: <reason>" when not applicable.

## Contract Checklist
- Template changes run `uv run python scripts/check_templates.py` and update snapshots or focused rendering tests as needed.
- Form/help/progress/error template changes check the producer data shape in Python code, docs examples, scaffold, and examples.
Comment thread docs/AGENTS.md
Comment on lines +21 to +24
- Public CLI, MCP, schema, scaffold, or verify changes update `docs/agent-quickstart.md` or `docs/testing.md`, or the PR explains why these docs are unaffected.
- New or changed command snippets use current flags, current import paths, and a runnable project shape.
- Structured error behavior changes include an agent-parseable example or a `no docs impact` note.
- Tagged code fences are covered by `uv run python scripts/check_docs_snippets.py`; untagged snippets must be intentionally illustrative.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants