forgebee-setup
Use when initializing ForgeBee on a new project — sets up CLAUDE.md, PM system, and project memory. For command-list questions, invoke the `forgebee-help` skill.
Best use case
forgebee-setup is best used when you need a repeatable AI agent workflow instead of a one-off prompt.
Use when initializing ForgeBee on a new project — sets up CLAUDE.md, PM system, and project memory. For command-list questions, invoke the `forgebee-help` skill.
Teams using forgebee-setup should expect a more consistent output, faster repeated execution, less prompt rewriting.
When to use this skill
- You want a reusable workflow that can be run more than once with consistent structure.
When not to use this skill
- You only need a quick one-off answer and do not need a reusable workflow.
- You cannot install or maintain the underlying files, dependencies, or repository context.
Installation
Claude Code / Cursor / Codex
Manual Installation
- Download SKILL.md from GitHub
- Place it in
.claude/skills/forgebee-setup/SKILL.mdinside your project - Restart your AI agent — it will auto-discover the skill
How forgebee-setup Compares
| Feature / Agent | forgebee-setup | Standard Approach |
|---|---|---|
| Platform Support | Not specified | Limited / Varies |
| Context Awareness | High | Baseline |
| Installation Complexity | Unknown | N/A |
Frequently Asked Questions
What does this skill do?
Use when initializing ForgeBee on a new project — sets up CLAUDE.md, PM system, and project memory. For command-list questions, invoke the `forgebee-help` skill.
Where can I find the source code?
You can find the source code on GitHub using the link provided at the top of the page.
SKILL.md Source
# ForgeBee Setup & Configuration
## Objective
Initialize ForgeBee for a project: CLAUDE.md, PM system, hook configuration. The user should have a working setup in under 5 minutes.
## Never
- Never overwrite an existing CLAUDE.md — merge or confirm with user first
- Never skip asking about the project's stack and conventions
- Never create directory structures without checking what already exists
Initialize and configure the ForgeBee development framework for any project. Walk the user through
setting up project memory, task tracking, and the project management system.
## First-Time Setup
When the user wants to initialize ForgeBee for their project:
1. **Create CLAUDE.md** from the template at `${CLAUDE_PLUGIN_ROOT}/skills/forgebee-setup/references/claude-md-template.md`
- Ask the user to fill in: stack, conventions, key components, common commands
- Write to the project root
2. **Create TASKS.md** if it doesn't exist
- Initialize with empty Active/Waiting On/Done sections
3. **Initialize PM system** by copying templates:
- `docs/pm/state.yaml` from `${CLAUDE_PLUGIN_ROOT}/templates/docs/pm/state.yaml`
- `docs/pm/index.md` from `${CLAUDE_PLUGIN_ROOT}/templates/docs/pm/index.md`
- `docs/pm/decisions.md` from `${CLAUDE_PLUGIN_ROOT}/templates/docs/pm/decisions.md`
- Create `docs/pm/features/` directory
- Create `docs/planning/briefs/`, `docs/planning/requirements/`, `docs/planning/stories/`
4. **Create session directories**
- `.claude/sessions/`
- `.claude/session-cache/context-backups/`
- `.claude/learnings/learnings.md`
## Command / Agent / Skill Reference
Do NOT maintain a command list here — it drifts the moment a surface is added or removed.
The authoritative inventory lives in two places:
- **`forgebee/INDEX.md`** — auto-generated by `scripts/build-index.js`, loaded on SessionStart. Canonical list of every command, agent, and skill with current counts. Never hand-edited.
- **`forgebee-help` skill** — invoke it when the user asks "what commands are available", "how do I use ForgeBee", or wants a cheat sheet. It reads INDEX.md and surfaces the relevant slice.
When a freshly-initialized user asks what they can run, point them at `forgebee-help` (or `forgebee/INDEX.md` directly). Setup ends once CLAUDE.md, the PM system, and session directories exist — discovery is the help skill's job, not this one's.
## Permission Mode Recommendation
Experienced users may prefer **auto mode** (fewer prompts — the permission-guard's Tier-0 blocklist still enforces in every mode, so dangerous commands stay blocked regardless):
```bash
claude --permission-mode auto
```
Auto mode uses Claude's AI safety classifier to approve safe operations automatically while blocking dangerous ones. ForgeBee's Tier 0 blocklist still fires before the classifier for maximum safety.
**Requirements:** Team/Enterprise plan, Anthropic API, Sonnet 4.6 or Opus 4.6.
For users who can't use auto mode, the default permission mode works well — ForgeBee's built-in allowlist handles common safe commands.
## Additional Resources
- **`references/claude-md-template.md`** — Full CLAUDE.md template for project memory
- **`references/pm-overview.md`** — Project management system documentationRelated Skills
forgebee-help
Use when the user asks "what commands are available", "how do I use ForgeBee", "what can ForgeBee do", or needs a quick reference. Reads forgebee/INDEX.md as canonical source.
strategy-skeptic
Use when /growth reaches the strategy debate phase — argues AGAINST marketing strategy, finds weak positioning, audience gaps, flawed assumptions.
strategy-judge
Use when /growth strategy debate needs adjudication — rules on each item after reading blind Advocate and Skeptic cases. Approve, block, or flag.
strategy-advocate
Use when /growth reaches the strategy debate phase — argues FOR marketing strategy artifacts, defends quality, feasibility, and effectiveness in blind debate.
review-wordpress
Use when reviewing WordPress plugin or theme code for WP coding standards (WPCS), security (nonces, sanitization, escaping), hook naming, text domains, or plugin architecture.
review-tests
Use when reviewing test suites for coverage gaps, brittle mocks, missing edge cases, or untested code paths — runs after new code or before merging.
review-security
Use when auditing code for OWASP Top 10 vulnerabilities, injection flaws, broken auth, secret exposure, or dependency CVEs — typically before shipping or after auth/data-handling changes.
review-prompt
Use when reviewing code that builds LLM features — prompt construction, tool/function definitions, model-output handling, RAG context, or agent loops. Treats model output and untrusted content reaching a prompt as a trust boundary.
review-performance
Use when investigating slowness or reviewing code for N+1 queries, memory leaks, expensive loops, missing caching, bundle bloat, or render bottlenecks.
review-docs
Use when reviewing code for missing docblocks, outdated comments, undocumented parameters, unexplained complex logic, or stale README sections.
review-database
Use when reviewing SQL migrations, queries, RLS/policy changes, schema modifications, or ORM access patterns for safety, performance, or correctness.
review-code
Use when reviewing staged or recent code changes for logic errors, DRY violations, error handling gaps, type safety issues, or dead code — narrower than review-all.