← Back to registry
Mapped Project Advanced by Agent Buildprint

Portable Personal Agent Chat OS

A phase-flow Buildprint that lets a downstream agent implement a portable personal agent chatbot OS while keeping deterministic proof separate from live/external claims.

Open manifest GitHub folder

methodology

agent-ready
01 Setup gate
02 Phase-flow execution
03 Runtime evidence ledger
04 Claim-safe validation

Executable packet spine

The files or runtime artifacts a fresh agent must read or produce before claiming progress.

  • BUILDPRINT.md
  • 01-questions.md
  • 02-project-setup.md
  • blueprint.yaml
  • 03-phases/phase-index.yaml
  • 03-phases/phase-flow.md
  • 03-phases/<active-phase>.md
  • 04-evaluation.md
  • 05-evidence/evidence-ledger.schema.json
  • .buildprint/phase-runs/<phase-id>/proof.md
  • .buildprint/evidence/evidence-ledger.jsonl

Proof gates

Checks that prevent a vague implementation from being reported as complete.

  • ✓Mapper selected-output structural check passes
  • ✓Executable packet starts from BUILDPRINT.md
  • ✓02-project-setup.md defines architecture, team model, execution authority, handoffs, AGENTS.md plan, quality gates, safety, assumptions, and phase start gate
  • ✓Every phase starts with the required implementation heading and routes through phase-flow artifacts
  • ✓Packaged evidence ledger is seed-only
  • ✓Runtime evidence belongs only in .buildprint/evidence/evidence-ledger.jsonl
  • ✓Evidence schema blocks blocked/missing/synthetic/partial rows from upgrading claims
  • ✓Publication copy prompt names the executable read order
  • ✓Legacy document sludge removed or reduced to compatibility spine files

Risks covered

Failure modes the Buildprint makes visible before an agent can hide them in “done”.

  • ⚠Overclaiming full Emperor Agent clone parity
  • ⚠Unsafe shell/filesystem/network/browser tools
  • ⚠Provider API drift
  • ⚠Context truncation losing user instructions
  • ⚠Skill injection bloat
  • ⚠MCP timeout/security failures
  • ⚠In-memory state being mistaken for durable persistence
  • ⚠Telemetry hidden from UI
  • ⚠Hosted auth, billing, publishing, media, or retrieval claims without proof

What this ships

01

BUILDPRINT.md canonical entry

02

01-questions.md setup questions

03

02-project-setup.md architecture and operating contract

04

blueprint.yaml machine contract

05

03-phases/phase-index.yaml phase index

06

03-phases/phase-flow.md orchestration protocol

07

Phase files for storage, runtime, tools/MCP, memory/team/telemetry, safety, and WebUI/API

08

04-evaluation.md evidence honesty gates

09

05-evidence/evidence-ledger.schema.json

10

Seed-only evidence ledger

Validation evidence

For newer phase-flow Buildprints, this names replay or outcome evidence. For older registry entries, it lists the included validation checks.

  • Mapper OS converted: Converted to executable-blueprint v5 phase-flow methodology with seed-only packaged evidence and honest runtime evidence rules.
  • Converted to executable-blueprint v5 with phase-run artifacts and review contracts.
  • Preserves personal agent OS behavior without copying source implementation or claiming exact clone parity.
  • Runtime evidence is separated from seed evidence and blocked proof cannot upgrade claims.
  • Shell, browser, network, filesystem writes, live providers, and real MCP are explicit approval/proof boundaries.

Copyable agent prompt

Manifest 
Use the Portable Personal Agent Chat OS Mapper OS packet. Bootstrap exact snapshots with agb start https://agent-buildprint.com/buildprints/portable-personal-agent-chat-os/package.json . Then read BUILDPRINT.md, 01-questions.md, 02-project-setup.md, blueprint.yaml, 03-phases/phase-index.yaml, 03-phases/phase-flow.md, the active phase file, 04-evaluation.md, and 05-evidence/evidence-ledger.jsonl. Treat BUILDPRINT.md as canonical. Run every phase through 03-phases/phase-flow.md, create .buildprint/phase-runs/<phase-id>/ artifacts, read .buildprint/next-agent.md and current project AGENTS.md before phase work, and append runtime proof only to .buildprint/evidence/evidence-ledger.jsonl. Build the clean-room personal agent chatbot OS with streaming chat, provider routing, tools, skills, MCP adapter, memory/compaction, subagent delegation, token telemetry, WebUI/API workbench, and safety-gated tool policy. Use deterministic providers/tools by default. Do not claim Emperor Agent clone, exact UI parity, live provider parity, real MCP parity, shell/browser/network safety, production auth, billing, publishing, or hosted SaaS unless separately validated.