Lead Engineer Agent

The builder of the engineering org in a zero-human AI startup. This agent is the highest-throughput producer — it receives prioritized tasks from the CTO, decomposes them into implementation plans, writes production-quality code, reviews all other engineering code, and coordinates with QA/Testing and Data Engineer to deliver complete, tested features every sprint.

Quick Start

Deploy the agent using OpenClaw with the ClawPack bundle:
```
clawpack install @agentebox/lead-engineer
```
Configure communication channels — the Lead Engineer needs to send/receive messages to the CTO (upstream) and QA/Testing, Data Engineer (downstream), plus DevOps/Infra (deployment coordination).
Set up the Remote Project Board — primary tracking for sprint tasks, bugs, and feature progress.
Connect the code repository — all code, PRs, reviews, and documentation flow through version control.
Configure cadences — daily standup (morning, 5 min), weekly sprint planning (Monday, 30 min), sprint execution (continuous).
Send the first directive — trigger sprint-planning with the CTO's initial task assignments and priorities.

Environment Variables

Variable	Description	Required
`REMOTE_PROJECT_ID`	Project ID on the Remote board	Yes
`CTO_AGENT_ID`	Session ID or label for the CTO agent	Yes
`QA_AGENT_ID`	Session ID or label for the QA/Testing agent	Yes
`DATA_ENGINEER_AGENT_ID`	Session ID or label for the Data Engineer agent	Yes
`DEVOPS_AGENT_ID`	Session ID or label for the DevOps/Infra agent	Yes
`CODE_REPO_URL`	Repository URL for the main codebase	Yes
`TEST_COVERAGE_TARGET`	Minimum test coverage on critical paths (default: 80%)	No

File Listing

File	Description
`SOUL.md`	Complete agent identity: behaviors, decision framework, communication protocols, boundaries, failure modes
`IDENTITY.md`	Quick-reference identity card (name, role, emoji)
`manifest.json`	Machine-readable configuration: skills, tools, cadences, autonomy levels
`README.md`	This file — setup guide and integration reference
`skills/feature-development/SKILL.md`	Task decomposition → implementation → self-review → PR submission
`skills/code-review/SKILL.md`	Structural + line-level review with actionable feedback and severity ratings
`skills/sprint-planning/SKILL.md`	Capacity calculation, task sequencing, dependency coordination, carryover management
`skills/technical-documentation/SKILL.md`	API docs, architecture records, onboarding guides, decision documentation

Architecture

CTO
     ↕ (assignments, design reviews, status reports)
Lead Engineer ──── 🛠️
     ├── QA / Testing (test suites, bug triage, regression testing)
     └── Data Engineer (pipelines, data quality, dashboards)

Also coordinates with:
     → DevOps / Infra (deployment, infrastructure requirements)

Framework Integration

OpenClaw (Native)

# openclaw.yaml
agent:
  name: lead-engineer
  soul: ./SOUL.md
  identity: ./IDENTITY.md
  skills:
    - ./skills/feature-development/
    - ./skills/code-review/
    - ./skills/sprint-planning/
    - ./skills/technical-documentation/
  heartbeat:
    interval: 30m
    file: ./HEARTBEAT.md

CrewAI

from crewai import Agent, Task, Crew

lead_eng = Agent(
    role="Lead Engineer",
    goal="Ship production-quality features every sprint while maintaining code quality and coordinating the engineering team",
    backstory=open("SOUL.md").read(),
    tools=[repo_tool, remote_board_tool, messaging_tool],
    verbose=True,
    allow_delegation=True
)

sprint_work = Task(
    description="Execute this sprint: decompose CTO assignments, implement features with tests, review QA and Data Engineer code, submit PRs for review",
    agent=lead_eng,
    expected_output="Completed features with tests, reviewed PRs, updated documentation"
)

crew = Crew(agents=[lead_eng], tasks=[sprint_work], verbose=True)
crew.kickoff()

Monitoring

The Lead Engineer is healthy when:

Features shipped meets or exceeds the sprint target
Test coverage stays ≥80% on critical paths
Bug escape rate stays below 5%
Estimation accuracy stays above 85% (estimated vs actual within ±15%)
Code reviews completed within SLA (4h for P0/P1, 24h for P2+)
Sprint plan published every Monday morning

Warning signs:

Features consistently not completing within sprint (velocity declining)
Bug escape rate rising (quality degradation under pressure)
Estimation accuracy below 70% for 3+ sprints (estimation drift)
Code reviews piling up (>48h in queue)
QA or Data Engineer reporting frequent coordination friction
Carryover tasks accumulating across sprints

Version History

Version	Date	Changes
1.0.0	2026-03-16	Initial creation