prompt-forge
Skill3 starsChanged 7 months agoMITView on GitHub
Use when a developer explicitly asks for help crafting or refining an execution prompt. Not triggered by general task requests — only when they specifically ask for prompt writing assistance, not task execution.
8 other projects use a version of this.
Adopted by
Most recently taken up by
- patneedimadhukar/prompt_forgetheir own version, first seen Nov 8, 2026
- mahik504/orchestra-workflowtheir own version, first seen Nov 1, 2026
- morteza-omidi/prompt-forgetheir own version, first seen Oct 25, 2026
- devpilotX/skillstheir own version, first seen Sep 22, 2026
- Lesile-Yin/agent-skillstheir own version, first seen Aug 26, 2026
- the-Nile-jones/claude-skillstheir own version, first seen Aug 3, 2026
Other versions8 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 8
- Lesile-Yin/agent-skills★ 2same name
- mturac/hermes-supercode-skills★ 2same name
- mahik504/orchestra-workflow★ 1same name
- morteza-omidi/prompt-forge★ 1same name
- devpilotX/skills★ 0same name
- hauntedstack/prompt-forge★ 0same name
- patneedimadhukar/prompt_forge★ 0same name
- the-Nile-jones/claude-skills★ 0same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
The file
We split this skill into the rules it gives an agent
1 rule. 0 are common, 1 is distinctive, and 0 are very specific to this project.
0
Common
Most of the 7 edited versions we read keep these.
1
Distinctive
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 72
When building prompts, read the relevant references:
“Common” is measured: kept by at least 6 in 10 of the 7 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Prompt Forge
A prompt refinement skill that extracts what a developer actually means — especially when they're too deep in a session to say it clearly. It investigates the codebase, researches the ecosystem, and surfaces the perspectives that fatigue makes you forget, then produces a grounded prompt that the developer hands to their execution tool.
Built on Anthropic's official prompting best practices. Designed to work as a plugin for Claude Code and Gemini CLI, alongside GSD and Superpowers.
The Cardinal Rule
Investigate freely — but never implement. Prompt Forge is a prompt writer, not a task executor. Its only job is to produce a refined, grounded prompt. See src/core/constraints.md for full boundary definition.
Core Workflow
The workflow has two grounding passes — one before asking questions, one after.
| Step | Module | What Happens |
| 1. Receive input | src/core/intent_parser.md | Read raw input, detect fatigue signals |
| 2. First grounding | src/core/intent_parser.md | CLAUDE.md → code analysis → web research |
| 3. Ask questions | src/core/intent_parser.md | 1-3 grounded, fatigue-friendly questions |
| 4. Second grounding | src/core/intent_parser.md | Targeted deep-dive from answers |
| 5. Apply lenses | src/core/prompt_builder.md | 9 perspective lenses (business, QA, architecture, UX, security, performance, DX, edge cases, migration) |
| 6. Produce output | src/core/prompt_builder.md | Intent breakdown + copy-paste prompt |
| 7. Scope check | src/core/constraints.md | Verify no implementation drift |
| 8. Deliver | src/utils/helpers.md | Present prompt, stay open for refinement |
Module Index
Core Logic
src/core/intent_parser.md— Steps 1-4: input parsing, grounding, clarifying questionssrc/core/prompt_builder.md— Steps 5-6: lens analysis, task classification, output generationsrc/core/modes.md— 5 compilation modes (build, audit, debug, research, optimize)src/core/constraints.md— The Cardinal Rule, scope boundaries, drift detection
LLM Adapters
src/adapters/claude.md— Claude/Anthropic formatting (XML tags, @ references)src/adapters/gemini.md— Gemini/Google formatting (MUST/MUST NOT, markdown)src/adapters/openai.md— OpenAI/GPT formatting (system/user split, few-shot)
Command
src/commands/prompt-forge.md— Slash command entrypoint and invocation guide
Helpers
src/utils/helpers.md— Tone, collaboration style, complexity adaptation, delivery protocol
Templates
prompts/templates/task-type-blueprints.md— 8 task-type prompt structuresprompts/templates/gsd-output-format.md— GSD-optimized output formatprompts/templates/superpowers-output-format.md— Superpowers-optimized output formatprompts/templates/context-file-template.md— Template for project context filesprompts/templates/anthropic-prompting-guide.md— Anthropic prompting best practices
Examples
prompts/examples/example-session.md— Full walkthrough: vague input → 3 output formats
Evaluation
evals/test_cases.md— 14 functional test casesevals/adversarial_cases.md— 15 boundary and failure mode testsevals/benchmark.md— Cross-model benchmark framework (8 cases)evals/scoring.md— Scoring rubric (clarity, constraints, structure, grounding, leakage)
Documentation
docs/architecture.md— System design and directory mapdocs/usage.md— Integration and usage guide
Reference Files
When building prompts, read the relevant references:
- Task-type blueprints — Primary reference on every invocation
- Anthropic prompting guide — Prompting principles: XML structuring, few-shot, chain-of-thought, grounding
- Context file template — For generating project context files
- GSD/Superpowers output formats — When plugins are detected
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.