Skip to content

Adrian/cicd and collab standards#16

Open
ZenAlexa wants to merge 19 commits intohakoniwaa:mainfrom
ZenAlexa:adrian/cicd-and-collab-standards
Open

Adrian/cicd and collab standards#16
ZenAlexa wants to merge 19 commits intohakoniwaa:mainfrom
ZenAlexa:adrian/cicd-and-collab-standards

Conversation

@ZenAlexa
Copy link
Collaborator

@ZenAlexa ZenAlexa commented Dec 11, 2025 •
edited
Loading

Summary

Add comprehensive CI/CD infrastructure and collaboration standards to establish consistent development workflow for the team.

Spec References

  • Implements: DR4 (Test Before Merge)
  • Creates: specs/CONTRIBUTING.md with new spec IDs (PR-001 to PR-003, COMMIT-001/002, REVIEW-001/002, ISSUE-001/002, SEC-001/002, RELEASE-001/002, COMM-001/002)
  • Updates: AGENTS.md (adds reference to CONTRIBUTING.md)

Type

  • feat: New feature
  • chore: Build/tooling changes

Changes

CI Pipeline (.github/workflows/ci.yml)

Job Description
rust-check Cross-platform (Linux/macOS/Windows), fmt + clippy + test + build
python-check Python 3.11/3.12, ruff format + lint + pytest with coverage
security cargo-audit, pip-audit, trufflehog secret scanning
specs Validates spec ID uniqueness across all spec files

Release Pipeline (.github/workflows/release.yml)

Stage Description
build Multi-platform binaries (x86_64 + aarch64 for Linux/macOS/Windows)
python-wheel Build Python wheel for agent
release Create GitHub Release with checksums
pypi Publish to PyPI

Collaboration Standards (specs/CONTRIBUTING.md)

Category Spec IDs Content
PR Workflow PR-001, PR-002, PR-003 Standard flow, template, review requirements
Commits COMMIT-001, COMMIT-002 Message format, hygiene rules
Reviews REVIEW-001, REVIEW-002 Checklist, etiquette
Issues ISSUE-001, ISSUE-002 Labels, templates
Security SEC-001, SEC-002 Secrets handling, dependency scanning
Releases RELEASE-001, RELEASE-002 Semver, checklist

Templates & Automation

  • PR template, Bug report template, Feature request template
  • Dependabot for weekly dependency updates

Checklist

  • Specs updated (new specs/CONTRIBUTING.md created)
  • Tests added/updated (N/A - infrastructure only)
  • Follows spec-driven development

Questions for Review

  1. CI Matrix: Is testing on all 3 platforms necessary for every PR?
  2. Review Requirements: 1 reviewer for code, 2 for specs - reasonable?
  3. Response Times: 24h security, 48h bugs/PRs - ok?
  4. Missing specs?: Any collaboration standards I missed?

kaminoguo and others added 19 commits December 6, 2025 13:01
@kaminoguo
docs: add PydanticAI as agent framework choice
93bd03d
@kaminoguo @claude
docs: add 2-tier LLM design and context compose flow
d5840eb 
🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude <noreply@anthropic.com>
@kaminoguo @claude
docs: add 3-layer storage architecture and team features
eee485a 
- Add individual vs team memory separation (squirrel.db vs group.db)
- Add 6 new team memory types (team_style, team_profile, shared_*, team_process)
- Add process memory type for audit/export
- Add team commands (sqrl share, sqrl export, sqrl team)
- Add team DB options (cloud/self-hosted/local)
- Merge v1.1 features into v1 scope
- Split v1 into Individual (free) and Team (paid) features

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
@kaminoguo @claude
docs: add complete technology stack decisions
c6c9d4b 
- Add Technology Stack table to README.md and DEVELOPMENT_PLAN.md
- IPC: JSON-RPC 2.0 (MCP-compatible)
- MCP SDK: rmcp (official Rust SDK)
- Build/Release: dist (cargo-dist) with Homebrew, MSI, installers
- Auto-update: axoupdater (dist's official updater)
- Python packaging: PyInstaller (bundled, zero user deps)
- Cloud sync: SQLite Session Extension + server_seq cursor
- Update installation section with MSI for Windows

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
@kaminoguo @claude
docs: add episode segmentation and memory lifecycle
f876a30 
- Episode segmentation: EXECUTION_TASK / PLANNING_DECISION / RESEARCH_LEARNING / DISCUSSION
- SUCCESS/FAILURE only for EXECUTION_TASK with evidence requirement
- Memory lifecycle: status (active/inactive/invalidated) + validity tracking
- Fact contradiction detection via semantic_key + LLM
- Soft delete mechanism (sqrl forget)
- Updated v1/v2 scope with limitations

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
@kaminoguo @claude
docs: add declarative keys and evidence source to v1 spec
83f0752 
- Add declarative key registry (project.* and user.* keys)
- Add evidence_source field for facts (success/failure/neutral/manual)
- Update memories table schema with key, value, support_count, last_seen_at
- Update episode ingestion output to include keyed facts
- Add context output structure for retrieval
- Update user_profile to structured identity table

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
@kaminoguo @claude
merge: resolve conflicts with main, keep API embeddings and declarati…
d628046 
…ve keys

Conflicts resolved:
- Keep API-based embeddings (text-embedding-3-small) over ONNX
- Keep declarative keys and evidence_source for facts
- Keep CLI selection and embedding config
- Keep segment-first design and memory lifecycle

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
@kaminoguo @claude
docs: add spec-driven development structure
a3b1dac 
- Add specs/ folder with Spec Kit pattern:
  - CONSTITUTION.md: project governance, principles P1-P5, rules DR1-DR5
  - ARCHITECTURE.md: system boundaries ARCH-001/002/003, data flows
  - SCHEMAS.md: database schemas SCHEMA-001 to SCHEMA-006
  - INTERFACES.md: IPC, MCP, CLI contracts with stable IDs
  - KEYS.md: declarative key registry KEY-P-*, KEY-U-*
  - PROMPTS.md: LLM prompts PROMPT-001 to PROMPT-005 with model tiers
  - DECISIONS.md: ADR log ADR-001 to ADR-009

- Add project-rules/ with .mdc format (Cursor + Claude compatible):
  - general.mdc, rust-daemon.mdc, python-agent.mdc, specs.mdc, testing.mdc

- Add devenv.nix for Nix development environment (ADR-006)
- Update CLAUDE.md to reference new structure
- Update README.md with devenv setup option

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
@kaminoguo @claude
merge: resolve conflicts with main, keep frustration detection
808c760 
🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
@kaminoguo @claude
docs: update agent instruction files with AI-first emphasis
65fb751 
- Add AI-First Development section (98% AI-coded)
- Emphasize declarative thinking and spec-driven approach
- Update date to 2025 Dec 9
- Reference spec IDs and project-rules/*.mdc

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
@kaminoguo @claude
fix: remove .mcp.json (contains secrets), update .cursorrules
a026be6 
- Remove .mcp.json from git (contains GitHub tokens)
- Remove Zone.Identifier Windows metadata files
- Add .mcp.json and *:Zone.Identifier to .gitignore
- Update .cursorrules with AI-first development emphasis

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
@kaminoguo @claude
docs: add frustration detection to development plan and example
2292ac6 
🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
@kaminoguo @claude
merge: resolve conflicts with main, keep AI-first emphasis
21a3b10 
🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
@kaminoguo @claude
refactor: consolidate agent instruction files
26060cd 
- AGENTS.md is now the single canonical source
- .claude/CLAUDE.md → symlink to AGENTS.md
- Delete GEMINI.md (configure Gemini to read AGENTS.md)
- Delete .cursorrules (deprecated by Cursor)
- Move project-rules/*.mdc → .cursor/rules/*.mdc
- Update specs/CONSTITUTION.md with new structure

This eliminates 4-file sync problem. Only AGENTS.md needs updating.

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
@kaminoguo
merge: resolve conflicts with main, keep consolidated agent files
1817e31 
- Keep AGENTS.md as canonical source with full content
- Keep .claude/CLAUDE.md as symlink → ../AGENTS.md
- Remove .cursorrules (deprecated, use .cursor/rules/*.mdc)
- Remove GEMINI.md (configure Gemini to read AGENTS.md)
@kaminoguo
docs: add AI workflow phases and pre-commit checklist to AGENTS.md
a4fd225 
Adopted from GitHub Spec Kit:
- 5-phase workflow: Specify → Clarify → Plan → Tasks → Implement
- NEEDS CLARIFICATION markers for unclear requirements
- Mandatory pre-commit checklist for spec-code consistency
@kaminoguo @claude
feat(hooks): add pre-commit hook for doc sync reminder
000e78d 
- Add .githooks/pre-commit that shows changed files before commit
- AI decides which docs to check based on file types
- Update AGENTS.md with guidance on which docs to check
- Hook does not block commits, only reminds

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
@kaminoguo @claude
chore: migrate useful content from DEVELOPMENT_PLAN.md, delete redund…
8b10858 
…ant files

- Add Technology Stack table to specs/ARCHITECTURE.md
- Add v2 Team/Cloud architecture to specs/DECISIONS.md
- Delete DEVELOPMENT_PLAN.md (superseded by specs/)
- Delete EXAMPLE.md (not needed)
- Delete .mcp.json:Zone.Identifier (Windows artifact)

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
@ZenAlexa @claude
feat(ci): add CI/CD pipeline and collaboration standards
0a1f909 
Add comprehensive CI/CD and collaboration infrastructure:

## CI Pipeline (.github/workflows/ci.yml)
- Cross-platform testing (Linux, macOS, Windows)
- Rust: fmt, clippy, test, build
- Python: ruff, pytest with coverage
- Security scanning (cargo-audit, pip-audit, trufflehog)
- Spec validation (ID uniqueness check)

## Release Pipeline (.github/workflows/release.yml)
- Multi-platform binary builds (x86_64 + aarch64)
- Python wheel packaging
- GitHub Release automation
- PyPI publishing

## Collaboration Standards (specs/CONTRIBUTING.md)
- PR workflow (PR-001, PR-002, PR-003)
- Commit standards (COMMIT-001, COMMIT-002)
- Code review guidelines (REVIEW-001, REVIEW-002)
- Issue management (ISSUE-001, ISSUE-002)
- Security policies (SEC-001, SEC-002)
- Release process (RELEASE-001, RELEASE-002)

## Templates
- Pull request template
- Bug report issue template
- Feature request issue template

## Automation
- Dependabot for dependency updates
- Markdown link checking

Implements: DR4 (Test Before Merge)

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
@kaminoguo
Copy link
Collaborator

kaminoguo commented Dec 11, 2025

Hey @ZenAlexa, heads up - we did a major v1 architecture redesign while this was in progress (ADR-010 in DECISIONS.md).

What changed:

  • New memory architecture: AI-primary, future-impact, declarative principles
  • New schema: memories, evidence, memory_metrics, episodes, guard_patterns tables
  • Single-stage Memory Writer replaces two-stage pipeline
  • New spec files: POLICY.md, V1_ARCHITECTURE_REDESIGN.md

The conflicts in ARCHITECTURE.md, CONSTITUTION.md, DECISIONS.md are from these changes.

Suggestion: Once lyrica/initial-setup merges to main, rebase your branch:

git fetch origin main
git rebase origin/main

Your CI/CD and CONTRIBUTING.md additions look good and don't conflict conceptually - just need the updated spec files as the base.

Re: your review questions:

  1. CI Matrix (3 platforms): Yes, worth it since we target Mac/Linux/Windows per ARCHITECTURE.md platform considerations
  2. Review requirements (1 code, 2 specs): Reasonable given specs are source of truth (DR6)
  3. Response times (24h security, 48h bugs): Looks good
  4. Missing specs: Consider adding a spec for the pre-commit hook workflow we added (.githooks/pre-commit for doc sync checks)

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

2 participants

@ZenAlexa @kaminoguo