forge
01Plugin Claude Code
Decide-then-code. Forge Mentor teaches every load-bearing decision, records your choice, and blocks code from moving past a decision you have not made yet.
22 15d ago A
tokens not measured
original MIT
forge
02MCP server Claude CodeCodexCursor +2
MCP server "forge" as configured in Hassaan146/forge-mentor. Runs ${CLAUDE_PLUGIN_ROOT}/server/forge_server.py with python. Needs 1 environment variable to run.
22 15d ago A
tokens not measured
original MIT
builder
03Agent
Writes the code for a decision that has already been recorded. Use after a decision exists in .claude/forge/decisions/, never before.
22 15d ago A 32 tokens
original MIT
planner
04Agent
Teaches a decision, presents options with a project-derived recommendation, and waits for the user. Also compiles phases and runs the explain-back gate. Never writes code.
22 15d ago A 36 tokens
original MIT
review-fixer
05Agent
Applies what CodeRabbit and Sourcery found on a pull request. Use after fetchreview has written .claude/forge/reviews/pr- .md.
22 15d ago A 38 tokens
original MIT
structurer
06Agent
Turns a user's free-text answer into a decision record in the schema. Use immediately after the user answers a question.
22 15d ago A 26 tokens
original MIT
add
07Command
Add a feature to a project Forge has already built, without re-asking what is already decided.
22 15d ago A 19 tokens
original MIT
mode
08Command
Change how much Forge settles on its own — pipeline, accept-edits, or auto.
22 15d ago A 17 tokens
original MIT
start
09Command
Start Forge Mentor in this project — connect accounts, agree permissions, and begin the foundation interrogation.
22 15d ago A 18 tokens
original MIT
status
10Command
Where this project stands, and what happens next.
22 15d ago A 9 tokens
original MIT
stop
11Command
Switch Forge off in this project, keeping every decision it recorded.
22 15d ago C 12 tokens
original MIT
update
12Command
Check whether a newer Forge is available, and show how to get it.
22 15d ago A 14 tokens
original MIT
PreToolUse
13Hook
Runs before the agent uses a tool for Write, Edit, NotebookEdit, Read, NotebookRead, Bash and Bash tool calls, executing governor.py, safety.py and gates.py via python (3 commands). From Hassaan146/forge-mentor.
22 15d ago A
tokens not measured
original MIT
PostToolUse
14Hook
Runs after a tool call finishes for mcp__forge__fetch_review, Bash, Write, Edit and NotebookEdit tool calls, executing reviewed.py and grounded.py via python (2 commands). From Hassaan146/forge-mentor.
22 15d ago A
tokens not measured
original MIT
SessionStart
15Hook
Runs when a session starts, executing forge_update.py, reviewed.py and companion.py via python with --hook (3 commands). From Hassaan146/forge-mentor.
22 15d ago A
tokens not measured
original MIT
Stop
16Hook
Runs when the agent finishes a response, executing presenter.py via python. From Hassaan146/forge-mentor.
22 15d ago A
tokens not measured
original MIT
Skill Claude CodeCodex
How Forge writes code, injected on every build step so the generated project stays coherent across phases. Use whenever writing or editing code in a Forge-governed project.
22 15d ago A 40 tokens
original MIT
Skill Claude CodeCodex
The gate at the end of a step — the user explains back what was built and why, before the step counts as finished. Use at the teach-back stage, after tests pass and the review is clean.
22 15d ago A 47 tokens
original MIT
Skill Claude CodeCodex
The minimums no decision can override. Loads at every Forge stage, without exception. Use when writing or reviewing any generated code, and when a user asks for something that would go below the floor.
22 15d ago A 44 tokens
original MIT
Skill Claude CodeCodex
How Forge teaches and questions a decision — teach the concept, give options with a project-derived recommendation, then wait. Use at interrogation and planning stages, before any code is written.
22 15d ago A 40 tokens
original MIT
Plugin Claude Code
Forge Mentor and related decide-then-code tooling for Claude Code.
22 16d ago A
tokens not measured