Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add richkuo/rk-skills --skill new-app-pipelinegit clone --depth 1 https://github.com/richkuo/rk-skillsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/richkuo/rk-skills/new-app-pipeline)<a href="https://agentmods.dev/skills/richkuo/rk-skills/new-app-pipeline"><img src="https://agentmods.dev/badge/skills/richkuo/rk-skills/new-app-pipeline/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/richkuo/rk-skills/new-app-pipeline"><img src="https://agentmods.dev/badge/skills/richkuo/rk-skills/new-app-pipeline.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00088 | $0.00742 |
| Opus 5 | $0.00044 | $0.00371 |
| Sonnet 5 | $0.00018 | $0.00148 |
| Haiku 4.5 | $0.00009 | $0.00074 |
Grade A, and why
new-app-pipeline scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 3d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 36 lines — stays where its author put it; the contents beside it link to each section on GitHub.
new-app-pipeline
The end-to-end process for starting a new app: capture the idea as a PRD, refine it by resolving every open question, break it into execution-ready GitHub issues, review the model/effort assignments, then run the milestone workflow. Each stage produces a durable artifact (PR, issues, workflow run) so the pipeline survives context resets — the conversation is never the state.
Stages
| # | Stage | Skill | Artifact | Checkpoint before next stage |
|---|---|---|---|---|
| 1 | Idea → PRD | app-prd |
PRD.md on a PR |
User iterates on the draft, in bursts |
| 2 | Resolve questions | prd-questions |
Updated PRD, empty Open Questions | User answered every batch |
| 3 | Merge the PRD PR | — | PRD on main | Explicit user go |
| 4 | Issues + milestones | prd-to-issues |
Milestones, complete scoped issues with Execution blocks | User reviews the breakdown table |
| 5 | Execution plan | execution-plan-review |
Revised Execution blocks | User settles the final table |
| 6 | Show the plan | milestoneplan |
Single per-issue plan table (complexity, dependencies, models, efforts, fableplan, first review) | User reviews the table (recommended, not required) |
| 7 | Run a milestone | milestone-workflow |
Workflow run → PRs → LGTMs | User approves the run plan (mandatory) |
Rules
- Never skip a checkpoint. Every stage boundary stops for the user; the pipeline is collaborative at the joints and autonomous inside them.
- Stages are re-enterable: the user can jump back ("actually keep Drizzle", "12 should be medium") at any point — apply the revision to the artifact, not just the chat.
- Stages 6–7 repeat per milestone (v0, then v1, …), chaining workflow invocations where cross-phase dependencies require earlier merges.
- Stage 6 is read-only and skippable — it never edits an issue. Skipping it means the first sign of a bad Execution stamp is an agent already running on it.
- Tech-stack debates during stage 1–2 (framework, ORM, hosting) are settled in chat and recorded in the PRD's Platforms & Technology table — the PRD carries the decision, the chat carries the reasoning.
- All the standing rules apply throughout: worktree + PR for every change, attribution footers,
github-issue-formatbefore filing, best-solution over cheap-solution.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 3d ago Changed df816633e73a
- 12d ago First seen · 36 lines · 88 tokens per session scan A 91949648e83a
new-app-pipeline is a skill published in the GitHub repository richkuo/rk-skills (49 stars, last pushed 2d ago), licensed MIT. It adds 88 tokens to every session and 742 once invoked, about $0.0004 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
contradiction-analysis
A structured way to analyze difficult problems by identifying competing forces, finding the main conflict, and deciding which side or issue has the greatest effect.
mass-line
A method for collecting feedback and facts from multiple people or sources, turning them into a plan, and checking the plan with those affected.
practice-cognition
A method for testing ideas in practice, learning from the results, and improving the next attempt through repeated cycles.
protracted-strategy
A long-term planning method for difficult work that cannot be completed quickly, divided into stages with steady progress over time.
concentrate-forces
A decision-making method for situations where several tasks compete for limited time, attention, computing power, or budget.
arming-thought
A Chinese-language routing guide built around checking facts before drawing conclusions. It helps select other methods for tasks such as investigation, decision-making, review, and long-term planning.