Team operating docs
A reusable Pi runtime for every project.
These docs bring installation, profiles, permissions, MCP, security, and workflows into one place. A new team member can find the right command, choose a safe mode, and understand the operating boundary without rebuilding the process in a prompt.
Production V1 · v1.6.0 public regression evidence
Measured, not asserted
Release-bound run production-v1-20260824T040017Z-05b7cf measured 108 sessions on GPT-5.6 Luna Medium at tag v1.6.0, exact commit 3bba8f0b3ff521bc2a355e1f6bef6d1bbdc09511, with Piagent compared against codex-cli.
Fixed-workload family fresh-token ratio
Tasks resolved
The primary fixed-workload family ratio is 0.3857, with a 95% confidence interval of 0.3073–0.4840. The point estimate is 61.43% lower, while the upper confidence bound supports the conservative statement that Piagent used at least 51.60% fewer fresh tokens on this predeclared production-v1 workload and exact release. Usage was exact for all 108 sessions and no retry occurred. See the benchmark methodology and evidence boundary.
Install with two commands
Requires Node.js >=22.19.0. Run these commands from the project you want to configure.
npm install -g @piagent/platform
piagent-setup
piagent-setup installs the pinned Pi Coding Agent host, installs the Pi package, initializes the current directory, and runs doctor. It also installs the MCP baseline and subagents; skip them with --no-mcp, --no-subagents, or --global-only. See the Quickstart for the complete flow.
Where to start
What’s new
Runtime truth, task trajectories, recovery, helper policy, and fail-closed completion receipts — including what runs automatically and what is opt-in.
Get startedQuickstart
Install, open a repository, onboard context, choose a profile and permission mode, run the first task, and verify before handoff.
Get startedPlatform scope
Why this repository exists alongside general coding agents, and which operating concerns it owns.
BuildProject profiles
Choose the operating profile for a repository: protected paths, verification commands, and granted capabilities.
MCPMCP
Four configuration layers, pinned presets, server management, authentication readiness, and repository approval gates.
OperateSecurity model
A policy enforcement layer, not an OS sandbox, with explicit claims about what is and is not controlled.
OperateProduction benchmark
Quality, safety, workflow, and token evidence from 108 sessions compared with the codex-cli surface.
Command catalog
Slash commands and terminal commands grouped by the job they perform.
Three documentation groups
Get started
For first-time users: installation, quickstart, and where the platform fits alongside other agent CLIs.
Build
Daily operation: profiles, permissions, commands, workflows, and a dedicated MCP section for its broader command surface.
Operate
For operators and reviewers: security boundaries, repository architecture, terminal tools, and the team operating standard.