Skip to content
Back to skills

Senmu Build Project

ASecurity

Initialize or reconcile project governance and AGENTS.md: working principles, owners, language and task-state rules. Not for ordinary execution or business-agent prompts.

  • 2 stars
  • 0 votes
  • 0 copies
  • 1 view
  • Added September 2, 2026
ai-agentsgoreactgitdocumentation

Security analysis

A100/100

Pro scans all 20 files and shows the line behind each finding

Scanned September 28, 2026

npx -y skills add SenMuShare/senmu-buildos --skill senmu-build-project --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Senmu Build Project?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Senmu Build Project
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/senmushare-senmu-build-project/badge)](https://www.skillsdirectory.com/skills/senmushare-senmu-build-project)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: senmu-build-project
description: "Initialize or reconcile project governance and AGENTS.md: working principles, owners, language and task-state rules. Not for ordinary execution or business-agent prompts."
---

# Project Governance

Own governance, authority and cross-domain boundaries—not other Skills. Ordinary work uses project entrypoints; continue for gaps, conflicts, evolution or explicit requests.

## Route by Outcome

- Create, assess, evolve governance: [Governance Instances](references/project-governance-instances-and-evolution.md).
- Staged takeover: [Takeover](references/established-project-takeover-governance.md).
- Lifecycle, capabilities, done: [Project Practice](references/project-lifecycle-guide.md).
- Roots, layout, document owners, maps: [Directories](references/project-directories-and-documentation.md).
- Low-cost task entry and maintenance: [Task Entry](references/task-entry-and-maintenance-economy.md).
- Reconcile AGENTS/host instructions and distill working principles: [Instruction Authoring](references/project-instruction-authoring.md). Preserve language, exceptions and code/non-code applicability; complete both tracks.
- Effective rules and conditional loading: [Standards Discovery](references/project-standard-discovery-and-on-demand-loading.md).
- Cross-stage task state: [Task State](references/task-execution-and-state-management.md).
- Handoffs and skill boundaries: [Adoption and Routing](references/project-adoption-handoff-and-scenario-routing.md).
- Actual G0-G4/gate decisions only: [Governance Levels](references/governance-levels-and-gates.md).

Read matching references only; Delivery owns Git.

New projects use [Initializer](scripts/init_project_governance.py); existing projects use read-only [Assessment](scripts/assess_project_governance.py). Output is bounded (`--verbose` expands registers), and proves neither acceptance, authority nor runtime behavior.

## Core Contract

- Establish roots, Git/subproject/release boundaries, entrypoints, owners, authority and non-goals. User intent outranks defaults within host permissions.
- Placement advice names the preferred owner/path and reason; advice alone does not authorize writes.
- Run `init_project_governance.py --mode plan-new` before authorized `initialize-new`; calibrate the draft and its language before adoption.
- Inventory existing projects read-only; confirm semantic ownership. Evolve original owners within authority, without default overwrites or parallel truth.
- Shape structure around actual capabilities and lifecycle, not speculative modules.
- Map capabilities to implementation, rules and checks under [Index Contract](references/project-standard-discovery-and-on-demand-loading.md#4-index-contract). AGENTS holds concise adopted principles, constraints, commands and routes—not manuals. Shared principles need not be project-unique.
- Use one Durable Task State Owner across stages; resuming a task does not reactivate Project.
- Plan/audit first; authorized repair or initialization continues in scope, audit-only remains read-only. Preserve exceptions; merge equivalents without repeated growth. Source edits do not prove installation or model behavior.

Handoff ownership with scope, facts, evidence, gaps, authority and recovery.

Files in this skill

  • SKILL.md3.2 KB
  • agents/openai.yaml419 B
  • assets/mature-project-governance/GOVERNANCE_CONTROL.template.json1.5 KB
  • assets/mature-project-governance/TAKEOVER_TASK.template.md2.2 KB
  • assets/project-governance-starter/.gitignore.template80 B
  • assets/project-governance-starter/AGENTS.template.md2.3 KB
  • assets/project-governance-starter/GOVERNANCE.template.md5.5 KB
  • assets/project-governance-starter/PROJECT_MAP.template.md3.6 KB
  • assets/project-governance-starter/README.template.md1.9 KB
  • assets/project-governance-starter/TASK.template.md2.8 KB
  • assets/project-governance-starter/TASK_REGISTER.template.md802 B
  • assets/project-governance-starter/WORKLOG.template.md616 B
  • references/任务执行与状态管理规范.md11.3 KB
  • references/成熟项目接管治理专项规范.md6.2 KB
  • references/治理强度分级与门禁规范.md10 KB
  • references/项目实践指南.md11.7 KB
  • references/项目治理实例与演进规范.md8.4 KB
  • references/项目目录与文档规范.md6 KB
  • references/项目落地移交与场景路由规范.md11 KB
  • references/项目规范发现与按需加载规范.md8.3 KB

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.