Prompts, workflows and more for agentic engineering
Python
2,260
278 commits
updated Oct 2, 2026
A collection of prompts for AI-assisted development with Claude Code.
https://www.youtube.com/watch?v=KVOZ9s1S9Gk&lc=UgzfwxvFjo6pKEyPo1R4AaABAg
Found value in these resources?
Buy me a coffee: https://coff.ee/wirasm
I spent a considerable amount of time creating these resources and prompts. If you find value in this project, please consider buying me a coffee to support my work.
Ready to move beyond toy demos to production-ready AI systems?
Book a workshop: https://www.rasmuswiding.com/
What you'll get:
Perfect for: Engineering teams, Product teams, and developers who want AI that actually works in production
Contact me directly at hello@rasmuswiding.com
Product Requirement Prompt (PRP) = PRD + curated codebase intelligence + agent/runbook
The minimum viable packet an AI needs to ship production-ready code on the first pass.
A PRP supplies an AI coding agent with everything it needs to deliver a vertical slice of working software—no more, no less.
A traditional PRD clarifies what the product must do and why customers need it, but deliberately avoids how it will be built.
A PRP keeps the goal and justification sections of a PRD yet adds AI-critical layers:
/plugin marketplace add Wirasm/PRPs-agentic-eng
/plugin install prp-core@prp-marketplace
Every skill is then available across all your projects as /prp-core:<name>, updated centrally.
# From your project root — copy individual skills (or the whole folder)
cp -r /path/to/PRPs-agentic-eng/.claude/skills/prp-plan .claude/skills/
git clone https://github.com/Wirasm/PRPs-agentic-eng.git
cd PRPs-agentic-eng
The .claude/skills/ directory contains the core PRP workflow as Agent Skills — also distributed as the prp-core plugin (invoke as /prp-core:<name>):
| Command | Description |
|---|---|
/prp-prd | Interactive PRD generator with implementation phases |
/prp-plan | Create implementation plan (from PRD or free-form input) |
/prp-implement | Execute a plan through validated commit and PR |
/prp-issue | Own an issue, plan, document, or idea through reviewed PR and green CI |
/prp-issue-contract | Create an issue or check its preconditions before agent automation |
/prp-prd-update | Maintain PRD phase status and delivery links |
/prp-companion | Visual HTML companion of a plan or review, beside its markdown |
| Command | Description |
|---|---|
/prp-debug | Diagnose a root cause and publish the evidence to the matching GitHub issue |
This skill requires explicit invocation because it can comment, label, or close another person's contribution.
| Command | Description |
|---|---|
/prp-maintainer-triage | Route a contributor PR or reported issue through its lightweight maintainer triage workflow |
| Command | Description |
|---|---|
/prp-commit | Smart commit with natural language file targeting |
/prp-pr | Create PR with template support |
/prp-review | Comprehensive PR code review |
| Command | Description |
|---|---|
/prp-loop | Autonomous pipeline: plan → implement with PR → review (review→fix loops until clean) |
/prp-orchestrate | Coordinate parallel workstreams running PRP skills in git worktrees, with review gates and merge sequencing |
| Command | Description |
|---|---|
/prp-deliver | Manual-only, top-level multi-context delivery experiment; excluded from distribution and orchestration |
| Command | Description |
|---|---|
/prp-codebase-question | Research how the codebase works using parallel agents |
/prp-research-team | Design a multi-agent research team and executable research plan |
/prp-meta-skill | Author new skills, or refactor fat skills into lean SKILL.md + references |
/prp-technical-writing | Write and edit developer documentation that is clear, concrete, and verified |
/prp-loop drives the full pipeline (plan → implement with commit and PR → review) headlessly, running one claude -p session per stage and looping review → fix until the PR review comes back clean. Progress is tracked in the project's PRP store at ~/.prp/<project-key>/state/prp-loop.state.json.
/prp-loop "add user authentication with JWT"
~/.prp/<project-key>/plans/<feature>.plan.md--max-cycles is reached# Run the full pipeline from a feature description
/prp-loop "add user authentication with JWT"
# Run through implementation and PR, then stop before review
/prp-loop "add user authentication with JWT" --until implement
# Resume a halted or in-progress loop
/prp-loop --resume
--until <stage> (plan | implement | pr | review | fix) halts once that stage completes; --until implement stops after the implementation is green, committed, and opened as a PR--max-cycles 3, --max-implement-iterations 10; base branch is auto-detected--validate "<cmd>" to give the loop an authoritative green check (exit 0 = pass)~/.prp/<project-key>/state/prp-loop.state.json; inspect or resume from there/prp-prd "user authentication system"
↓
Creates PRD with Implementation Phases table
↓
/prp-plan ~/.prp/<project-key>/prds/user-auth.prd.md
↓
Auto-selects next pending phase, creates plan
↓
/prp-implement ~/.prp/<project-key>/plans/user-auth-phase-1.plan.md
↓
Implements and validates, then commits and opens a PR
↓
Links the plan, report, and PR back to the source PRD
↓
/prp-review
/prp-plan "add pagination to the API"
↓
Creates implementation plan from description
↓
/prp-implement ~/.prp/<project-key>/plans/add-pagination.plan.md
/prp-issue 123
↓
Creates or resolves the plan, implements it, commits, and opens the PR
↓
Publishes the complete agent review on the PR
↓
Fixes or disproves review findings, re-reviews, and waits for green CI
Artifacts and runtime state are stored outside the repository in a per-project directory:
~/.prp/<project-key>/
├── project.json # Canonical project path and name
├── prds/ # Product requirement documents
├── plans/ # Implementation plans
├── research/ # Codebase research
├── research-plans/ # Multi-agent research plans
├── reports/ # Implementation reports
├── reviews/ # Human-readable PR reviews
├── debug/ # Root-cause analysis reports
├── orchestration/ # Parallel-workstream run files
└── state/ # Loop state, logs, and hook sentinels
The project key is derived from the canonical main-checkout path, so all linked worktrees share the same store. Set PRP_HOME to override the default ~/.prp root.
PRDs include an Implementation Phases table for tracking progress:
| # | Phase | Description | Status | Parallel | Depends | Plan | Report | PR |
| --- | ----- | ----------- | ----------- | -------- | ------- | ------ | ------ | ------ |
| 1 | Auth | User login | complete | - | - | [link] | [link] | [link] |
| 2 | API | Endpoints | in-progress | - | 1 | [link] | [link] | [link] |
| 3 | UI | Frontend | pending | with 4 | 2 | - | - | - |
| 4 | Tests | Test suite | pending | with 3 | 2 | - | - | - |
pending → in-progress → completeyour-project/
├── .claude/
│ ├── skills/ # PRP skills (or install the prp-core plugin)
│ └── agents/ # Custom subagents
├── CLAUDE.md # Project-specific guidelines
└── src/ # Your source code
When PRD phases can run in parallel:
# Phase 3 and 4 can run concurrently
git worktree add -b phase-3-ui ../project-phase-3
git worktree add -b phase-4-tests ../project-phase-4
# Run Claude in each
cd ../project-phase-3 && claude
cd ../project-phase-4 && claude
The previous slash-command generation is preserved in old-prp-commands/ for reference only. That includes its PRP templates (old-prp-commands/PRPs/templates/ — prp_base.md, prp_story_task.md, prp_planning.md) and curated AI documentation (old-prp-commands/PRPs/ai_docs/). The current skills carry their templates inline or in their own templates/ directories.
MIT — see LICENSE.
I spent a considerable amount of time creating these resources and prompts. If you find value in this project, please consider buying me a coffee to support my work.
Buy me a coffee: https://coff.ee/wirasm
The goal is one-pass implementation success through comprehensive context.
970 followers · starred Aug 2025
29 followers · starred Aug 2025
438 followers · starred Jul 2025
137 followers · starred Jul 2025
Prompts, workflows and more for agentic engineering
Python
2,260
278 commits
updated Oct 2, 2026
A collection of prompts for AI-assisted development with Claude Code.
https://www.youtube.com/watch?v=KVOZ9s1S9Gk&lc=UgzfwxvFjo6pKEyPo1R4AaABAg
Found value in these resources?
Buy me a coffee: https://coff.ee/wirasm
I spent a considerable amount of time creating these resources and prompts. If you find value in this project, please consider buying me a coffee to support my work.
Ready to move beyond toy demos to production-ready AI systems?
Book a workshop: https://www.rasmuswiding.com/
What you'll get:
Perfect for: Engineering teams, Product teams, and developers who want AI that actually works in production
Contact me directly at hello@rasmuswiding.com
Product Requirement Prompt (PRP) = PRD + curated codebase intelligence + agent/runbook
The minimum viable packet an AI needs to ship production-ready code on the first pass.
A PRP supplies an AI coding agent with everything it needs to deliver a vertical slice of working software—no more, no less.
A traditional PRD clarifies what the product must do and why customers need it, but deliberately avoids how it will be built.
A PRP keeps the goal and justification sections of a PRD yet adds AI-critical layers:
/plugin marketplace add Wirasm/PRPs-agentic-eng
/plugin install prp-core@prp-marketplace
Every skill is then available across all your projects as /prp-core:<name>, updated centrally.
# From your project root — copy individual skills (or the whole folder)
cp -r /path/to/PRPs-agentic-eng/.claude/skills/prp-plan .claude/skills/
git clone https://github.com/Wirasm/PRPs-agentic-eng.git
cd PRPs-agentic-eng
The .claude/skills/ directory contains the core PRP workflow as Agent Skills — also distributed as the prp-core plugin (invoke as /prp-core:<name>):
| Command | Description |
|---|---|
/prp-prd | Interactive PRD generator with implementation phases |
/prp-plan | Create implementation plan (from PRD or free-form input) |
/prp-implement | Execute a plan through validated commit and PR |
/prp-issue | Own an issue, plan, document, or idea through reviewed PR and green CI |
/prp-issue-contract | Create an issue or check its preconditions before agent automation |
/prp-prd-update | Maintain PRD phase status and delivery links |
/prp-companion | Visual HTML companion of a plan or review, beside its markdown |
| Command | Description |
|---|---|
/prp-debug | Diagnose a root cause and publish the evidence to the matching GitHub issue |
This skill requires explicit invocation because it can comment, label, or close another person's contribution.
| Command | Description |
|---|---|
/prp-maintainer-triage | Route a contributor PR or reported issue through its lightweight maintainer triage workflow |
| Command | Description |
|---|---|
/prp-commit | Smart commit with natural language file targeting |
/prp-pr | Create PR with template support |
/prp-review | Comprehensive PR code review |
| Command | Description |
|---|---|
/prp-loop | Autonomous pipeline: plan → implement with PR → review (review→fix loops until clean) |
/prp-orchestrate | Coordinate parallel workstreams running PRP skills in git worktrees, with review gates and merge sequencing |
| Command | Description |
|---|---|
/prp-deliver | Manual-only, top-level multi-context delivery experiment; excluded from distribution and orchestration |
| Command | Description |
|---|---|
/prp-codebase-question | Research how the codebase works using parallel agents |
/prp-research-team | Design a multi-agent research team and executable research plan |
/prp-meta-skill | Author new skills, or refactor fat skills into lean SKILL.md + references |
/prp-technical-writing | Write and edit developer documentation that is clear, concrete, and verified |
/prp-loop drives the full pipeline (plan → implement with commit and PR → review) headlessly, running one claude -p session per stage and looping review → fix until the PR review comes back clean. Progress is tracked in the project's PRP store at ~/.prp/<project-key>/state/prp-loop.state.json.
/prp-loop "add user authentication with JWT"
~/.prp/<project-key>/plans/<feature>.plan.md--max-cycles is reached# Run the full pipeline from a feature description
/prp-loop "add user authentication with JWT"
# Run through implementation and PR, then stop before review
/prp-loop "add user authentication with JWT" --until implement
# Resume a halted or in-progress loop
/prp-loop --resume
--until <stage> (plan | implement | pr | review | fix) halts once that stage completes; --until implement stops after the implementation is green, committed, and opened as a PR--max-cycles 3, --max-implement-iterations 10; base branch is auto-detected--validate "<cmd>" to give the loop an authoritative green check (exit 0 = pass)~/.prp/<project-key>/state/prp-loop.state.json; inspect or resume from there/prp-prd "user authentication system"
↓
Creates PRD with Implementation Phases table
↓
/prp-plan ~/.prp/<project-key>/prds/user-auth.prd.md
↓
Auto-selects next pending phase, creates plan
↓
/prp-implement ~/.prp/<project-key>/plans/user-auth-phase-1.plan.md
↓
Implements and validates, then commits and opens a PR
↓
Links the plan, report, and PR back to the source PRD
↓
/prp-review
/prp-plan "add pagination to the API"
↓
Creates implementation plan from description
↓
/prp-implement ~/.prp/<project-key>/plans/add-pagination.plan.md
/prp-issue 123
↓
Creates or resolves the plan, implements it, commits, and opens the PR
↓
Publishes the complete agent review on the PR
↓
Fixes or disproves review findings, re-reviews, and waits for green CI
Artifacts and runtime state are stored outside the repository in a per-project directory:
~/.prp/<project-key>/
├── project.json # Canonical project path and name
├── prds/ # Product requirement documents
├── plans/ # Implementation plans
├── research/ # Codebase research
├── research-plans/ # Multi-agent research plans
├── reports/ # Implementation reports
├── reviews/ # Human-readable PR reviews
├── debug/ # Root-cause analysis reports
├── orchestration/ # Parallel-workstream run files
└── state/ # Loop state, logs, and hook sentinels
The project key is derived from the canonical main-checkout path, so all linked worktrees share the same store. Set PRP_HOME to override the default ~/.prp root.
PRDs include an Implementation Phases table for tracking progress:
| # | Phase | Description | Status | Parallel | Depends | Plan | Report | PR |
| --- | ----- | ----------- | ----------- | -------- | ------- | ------ | ------ | ------ |
| 1 | Auth | User login | complete | - | - | [link] | [link] | [link] |
| 2 | API | Endpoints | in-progress | - | 1 | [link] | [link] | [link] |
| 3 | UI | Frontend | pending | with 4 | 2 | - | - | - |
| 4 | Tests | Test suite | pending | with 3 | 2 | - | - | - |
pending → in-progress → completeyour-project/
├── .claude/
│ ├── skills/ # PRP skills (or install the prp-core plugin)
│ └── agents/ # Custom subagents
├── CLAUDE.md # Project-specific guidelines
└── src/ # Your source code
When PRD phases can run in parallel:
# Phase 3 and 4 can run concurrently
git worktree add -b phase-3-ui ../project-phase-3
git worktree add -b phase-4-tests ../project-phase-4
# Run Claude in each
cd ../project-phase-3 && claude
cd ../project-phase-4 && claude
The previous slash-command generation is preserved in old-prp-commands/ for reference only. That includes its PRP templates (old-prp-commands/PRPs/templates/ — prp_base.md, prp_story_task.md, prp_planning.md) and curated AI documentation (old-prp-commands/PRPs/ai_docs/). The current skills carry their templates inline or in their own templates/ directories.
MIT — see LICENSE.
I spent a considerable amount of time creating these resources and prompts. If you find value in this project, please consider buying me a coffee to support my work.
Buy me a coffee: https://coff.ee/wirasm
The goal is one-pass implementation success through comprehensive context.
970 followers · starred Aug 2025
29 followers · starred Aug 2025
438 followers · starred Jul 2025
137 followers · starred Jul 2025