← Back to Skills Marketplace
105
Downloads
0
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install projitive
Description
Projitive is an MCP-first governance skill for agent-driven delivery. Use this before changing task states or writing governance artifacts. Core flow: taskNe...
Usage Guidance
Before installing or enabling this skill: (1) Understand it tells the agent to run `npm install -g @projitive/mcp@latest` and to initialize/modify a .projitive governance tree in your repo without asking. That means network/package install + automatic file writes. (2) Verify the npm package and its GitHub repo (https://github.com/yinxulai/projitive) and review its code and maintainers; prefer pinning to a specific vetted version instead of `latest`. (3) If you allow it, run the package in an isolated environment (container or sandbox) first, or restrict the agent so it must request explicit permission before creating/modifying files. (4) If you do not trust the external package or automatic repo writes, do not enable autonomous invocation and require manual approval for any projectInit or file changes.
Capability Analysis
Type: OpenClaw Skill
Name: projitive
Version: 2.1.3
The skill bundle instructs the agent to perform a global installation of an external package (`npm install -g @projitive/mcp`) and explicitly directs it to initialize project governance (`projectInit`) without seeking user permission in SKILL.md. While these actions are consistent with the stated goal of project management, the automated installation of remote code and the instruction to bypass user consent for system-level changes are high-risk behaviors that could be leveraged for unauthorized modifications.
Capability Assessment
Purpose & Capability
Name/description (governance/MCP-first) align with the SKILL.md: it focuses on discovery, context, task lifecycle, and writing governance artifacts. However, the SKILL.md requires using @projitive/mcp (npm) even though the skill metadata declared no required binaries or install spec — a mismatch between declared requirements and runtime instructions.
Instruction Scope
The instructions tell the agent to locate or create a .projitive repo root, call projectInit immediately ("Do NOT ask the user"), and read/write governance files (README.md, roadmap.md, tasks.md, designs/, reports/, templates/). Those file reads/writes are coherent for a governance skill, but the explicit directive to initialize and modify repository files autonomously (without user confirmation) expands the agent's scope and risk surface.
Install Mechanism
SKILL.md directs: `npm install -g @projitive/mcp@latest`. The package comes from the public npm ecosystem (not bundled or declared in the skill metadata). Requiring a global, always-latest install is a supply-chain and stability risk; the skill package did not include an install spec or declare this dependency in metadata, which is an inconsistency.
Credentials
The skill requests no environment variables, credentials, or config paths. All actions are limited to repository files and the external npm package; there is no direct request for secrets or unrelated credentials.
Persistence & Privilege
always:false (normal) and autonomous invocation is allowed (platform default). The skill instructs writing and modifying repository governance files and running projectInit; while these are scoped to the project, automatic, unprompted writes are functionally privileged and should be confirmed by the user or sandboxed.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install projitive - After installation, invoke the skill by name or use
/projitive - Provide required inputs per the skill's parameter spec and get structured output
Version History
v2.1.3
Initial release of projitive skill for MCP-first governance workflows:
- Introduces agent-driven delivery flow: taskNext → taskContext → execute → verify → taskNext.
- Requires usage of MCP methods for all governance discovery, context, and task advancement.
- Enforces mandatory governance files and evidence-based task state changes.
- Outlines clear execution loop with explicit task and evidence update rules.
- Provides fallback procedure for non-MCP environments.
Metadata
Frequently Asked Questions
What is Projitive?
Projitive is an MCP-first governance skill for agent-driven delivery. Use this before changing task states or writing governance artifacts. Core flow: taskNe... It is an AI Agent Skill for Claude Code / OpenClaw, with 105 downloads so far.
How do I install Projitive?
Run "/install projitive" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Projitive free?
Yes, Projitive is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does Projitive support?
Projitive is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Projitive?
It is built and maintained by Alain (@yinxulai); the current version is v2.1.3.
More Skills