scripts/`) are executable code for API calls, data fetching, report generation
- References (`references/`) are documentation for Claude to read and apply
- Scripts handle I/O; references handle knowledge
**Output Generation:**
- Skills
release after the replacement ships, warn with the replacement, then delete. Only surfaces documented `experimental` may break without that window.
- **Internal code:** Go packages, `web/`, `@compozy/ui`, specs, RFCs, and `.compozy/tasks
Claude Code: Best Practices for Effective Collaboration
This document outlines best practices for working with Claude Code to ensure efficient and successful software development tasks.
## Task Management
For complex
spreadsheets
- Web scraping or API development (unless specifically for spreadsheet data import)
- Non-spreadsheet document processing (use appropriate specialized agents)
- General system administration tasks
- Tasks that don't involve spreadsheet
capital, exchange-aligned
auto square-off) and support Telegram alerts.
Repository: https://github.com/marketcalls/openalgo
Documentation: https://docs.openalgo.in
## Documentation Map
All project documentation lives under `docs/` as markdown (the single source
plugin/plugin.json` is an added file in the diff) must ship the tiered submission documents per the matrix in `templates/skill-docs/README.md` (micro → `docs/PRD.md`; standard → + `docs/ADR.md`; pack, 2+ skills → + `docs/ONE-PAGER.md`; `CFO-ONE-PAGER.md` stays
context keys so asymmetric models (Cohere, Voyage, nomic-embed) can use different kwargs for document vs. query embedding.
Two install flavors: `[full]` bundles `sentence-transformers` (local inference); slim uses only
through. ⚠ Two full runs contend on the same `~/.code-index` process-lock scopes, the documented cause of .261's 47m outlier, so the 3.13 reproduce runs AFTER the local suite, never
/memory` to inspect loaded files. Root CLAUDE.md survives `/compact`.
## Claude Code Official Documentation
When working on Claude Code features (hooks, skills, subagents, MCP servers), use the `claude-code-guide` subagent
Model, Provider, Account, or Billing Source.
## Implementation Principles
- Inspect related implementations, tests, contracts, and documentation before making changes. Prefer established repository patterns and public APIs over parallel implementations.
- Reuse code
test section of @.github/workflows/ci.yml)
### Comments
- Comments MUST only be used to:
1. Document a function
2. Explain the WHY of a piece of code
3. Explain a particularly complicated piece
allowed:
- Quick exploration and prototyping
- Small fixes that don't need TDD workflow
- Documentation updates
### TDD Mode (Opt-In)
TDD is activated by `/work-on-issue`, which writes a `.tdd-session.json` marker.
When
cannot be executed from Windows.
When uncertain, make the smallest isolated change and explicitly document what still requires physical verification.
---
## Filesystem and paths
Never hardcode operating-system-specific paths
before** making any edits.
4. Wait for the user to explicitly accept before proceeding.
### Documentation Maintenance
Before working on code in `ace/`, read `docs/design/ACE_ARCHITECTURE.md` to understand the current architecture.
Before
boundary conditions and error states
5. **User experience** - Accessibility, performance, and usability considerations
6. **Documentation** - Add comments for complex logic, but prefer self-documenting code
---
Most formatting and common issues
binary's own flag table and is always current. When
`--help` and a document disagree, `--help` wins and the document is a bug.
## Build
```bash
cmake -S . -B build && cmake
A file Claude Code reads at the start of every session. It holds the commands, conventions and warnings the agent needs for this project.
Where does it go?
At the repository root. Claude Code also reads CLAUDE.md files in subdirectories when it works there.
What should it contain?
Build and test commands, the project's layout, conventions that aren't obvious from the code, and mistakes to avoid. Short files tend to work better than long ones.
CLAUDE.md or AGENTS.md?
Claude Code reads CLAUDE.md; most other agents read AGENTS.md. Many projects keep one and point the other at it.