Failures route through
`on_child_task_failed`.
### Format engine plugin system
To add a document format: subclass `FormatEngine` (`backend/format_engines/base.py`), set
`SUPPORTED_EXTENSIONS` / `FORMAT_NAME` / `FORMAT_DESCRIPTION`, implement `parse()` returning
`{format
provides Playwright-based browser automation.
It exposes **tools only** — no prompts or resources. Documentation ships with Next.js itself (`node_modules/next/dist/docs/`), and upgrade/Cache Components workflows are distributed as agent skills
with code in this repository.
## Repository Overview
This is a skills repository for Office document manipulation (PPTX, DOCX, XLSX, PDF). Each skill provides workflows, scripts, and documentation for working with
behavior, include scaffold-specific tests. If a change alters
provider rendering, include drift detection.
## Documentation Expectations
Update docs in the same change when behavior changes:
- `README.md` for user workflows
dependency, so no need to tell users to install it separately
### 4. Documentation Clarity
- Clearly distinguish between shell commands and Claude prompts
- Use `>` prefix for Claude prompts to avoid confusion
ALWAYS prefer editing an existing file to creating a new one
- NEVER proactively create documentation files (\*.md) or README files unless explicitly requested
- NEVER save working files, text/mds, or tests
long-running release or
development work.
- It is the only living work-status document. Update it after material
cross-session work; do not create parallel progress or dev-log files
spec:update` (see [Regenerating the OpenAPI clients](#regenerating-the-openapi-clients))
- Create markdown documentation files unless explicitly requested
- Mark Docker tests with `#[ignore]` -- they MUST skip gracefully at runtime instead
site's content model), see [CONTEXT.md](./CONTEXT.md).
For feature cards, feature pages, product documentation, and blog posts, follow the scoped [Content Authoring Guide](./CONTENT-AUTHORING.md
expires_at is not supported for this client"`) if they try it — don't document or suggest it to general agents.
### payment-methods command
- `payment-methods update --nickname ` requires
which prompt), and lets you blame, log, and inspect steps across sessions.
This document is the project's source of truth for context, vocabulary, and architectural decisions
source code (DO NOT MODIFY)
│ ├── multi_agent_supervisor.py
│ ├── prompts.py
│ ├── research_agent.py
│ ├── research_agent_mcp.py
│ ├── state_*.py
│ └── utils.py
└── README.md # Comprehensive documentation
```
## 🚨 Important Development Workflow
**The notebooks in `notebooks/` are the source of truth and should
Claude Code when working in this repository.
## Repository Overview
This is the MongoDB Documentation Monorepo. It combines three areas in one repo:
- `content/` — RST documentation source files for all MongoDB
Testing Platform
Grove is the infrastructure for creating **tested, snippeted code examples** for MongoDB documentation. Writers create example files with Bluehawk markup, test files that verify the examples work
types (metric, chart, table, text), filters, query templating (Jinja), grid layout, and CLI usage.
## Documentation
After every code change, scan `docs/` for areas the change touches and update them
fields, severity, categories, lifecycle) is owned by `skills/skill-health/SKILL.md`; the end-to-end loop is documented in `docs/CORE.md`. Only active once `INDEX.md` exists.
- **`memory/skill-health/`** — Per-run quality scores the health loop
English is the default language for all versioned repository content, including source comments,
JSDoc, documentation, changelog entries, user-facing strings, commit messages, and pull requests.
- Never add Portuguese prose
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.