P7
Run supervised sub-task execution with an implementation plan, code, and a three-question self-review when your agent stack spawns a P7 worker under P8.
Overview
P7 is an agent skill most often used in Build (also Ship review workflows) that executes sub-tasks under P8 supervision with plan-first implementation, code delivery, and a three-question self-review ending in [P7-COMPLE
Install
npx skills add https://github.com/tanweai/pua --skill p7What is this skill?
- Solution-driven flow: design plan and impact analysis before writing code
- Delivers implementation plan, code, and a mandated three-question self-review
- Reports completion through the [P7-COMPLETION] handoff for P8 supervision
- Inherits PUA core red lines and narration protocol from the parent /pua skill
- Detailed behavior defined in ../pua/references/p7-protocol.md
- 3-question self-review
- [P7-COMPLETION] delivery contract
Adoption & trust: 1.3k installs on skills.sh; 18k GitHub stars; 3/3 security scanners passed (skills.sh audits); trending (+100% hot-view momentum).
What problem does it solve?
You have a decomposed sub-task from a supervisor agent but need disciplined plan-then-code execution with a standard completion signal instead of unstructured coding.
Who is it for?
Multi-agent PUA setups where P8 assigns bounded implementation work and you want repeatable senior-engineer execution with explicit review and completion markers.
Skip if: Greenfield brainstorming without P8 context, tasks that only need a one-line fix with no plan gate, or teams not using the tanweai PUA parent skill and p7-protocol reference.
When should I use this skill?
User says 'P7模式' or '方案驱动', or when spawned as a sub-task executor by P8.
What do I get? / Deliverables
You get an implementation plan, working code, and a three-question self-review packaged for P8 via [P7-COMPLETION], ready for upstream merge or further orchestration.
- Implementation plan with impact analysis
- Implementation code
- Three-question self-review and [P7-COMPLETION] message
Recommended Skills
Journey fit
Spans multiple journey phases - primary shelf plus alternate fits below.
Canonical shelf is Build → agent-tooling because P7 is the dedicated executor role that turns approved plans into code inside a multi-agent PUA workflow. Fits agent-tooling as a procedural sub-agent protocol (plan → implement → [P7-COMPLETION]), not a one-off IDE snippet.
Where it fits
P8 assigns an API endpoint sub-task and P7 delivers plan, code, and self-review before [P7-COMPLETION].
A UI slice is handed to P7 with impact analysis on shared components before implementation.
The three-question self-review acts as a lightweight gate before P8 accepts the sub-task merge.
How it compares
Use as a governed sub-agent executor in a P8→P7 stack instead of a single chat session doing plan and code in one undifferentiated role.
Common Questions / FAQ
Who is p7 for?
Builders running tanweai’s PUA multi-agent workflow who need a P7 sub-executor that plans, codes, and self-reviews under P8 supervision.
When should I use p7?
During Build when P8 spawns implementation work; also in Ship-style review passes when you want the three-question self-review before marking a sub-task complete; invoke when the user says P7模式 or 方案驱动.
Is p7 safe to install?
Review the Security Audits panel on this Prism page and read p7-protocol.md and the parent PUA skill before granting your agent write and shell access during execution.
Workflow Chain
Requires first: pua
SKILL.md
READMESKILL.md - P7
# PUA P7 骨干 — 方案驱动执行 > 在 P8 管理下执行子任务。先设计方案 + 影响分析，再实施编码，完成后三问自审查。 详细协议见 `../pua/references/p7-protocol.md`。加载后按协议执行。 核心行为遵循 `/pua` 核心 skill 的三条红线和旁白协议。