Installable skillMIT

implementation-execution

Execute development tasks to build features, producing code, tests, and configuration artifacts that satisfy specification requirements and comply with constitution standards.

BY a5c-aiUPDATED Oct 28, 2026
INTENDED FOR◈OpenClawInstall in a listed agent.

instructions.md

---
name: implementation-execution-4916dfc250c9
description: Execute development tasks to build features, producing code, tests, and configuration artifacts that satisfy specification requirements and comply with constitution standards.
allowed-tools: Read, Write, Edit, Bash, Grep, Glob, Agent, AskUserQuestion
graph:
  domains: [domain:software-engineering]
  skillAreas: [skill-area:agentic-loops, skill-area:orchestration-loop]
  workflows: [workflow:feature-development]
  topics: [topic:developer-experience]
  roles: [role:tech-lead, role:backend-engineer]
---

- When implementing feature code against a specification
- When running per-task tests during implementation
- When validating implementation against specification requirements

## Key Principle

Every line of code must trace back to a specification requirement. Code quality must meet constitution standards. Tests must verify acceptance criteria from the task definition.

## Process

1. **Prepare environment** - Scaffold directories and configuration
2. **Execute independent tasks** - Parallel execution for tasks without dependencies
3. **Execute dependent tasks** - Sequential execution respecting dependency graph
4. **Run per-task tests** - Verify each task's acceptance criteria
5. **Validate against spec** - Check requirement coverage and story satisfaction
6. **Run integration tests** - Cross-component testing
7. **Human review** - Approve implementation before quality checklist

## Tool Use

Invoke via babysitter process: `methodologies/spec-kit/spec-kit-implementation`
Full pipeline: `methodologies/spec-kit/spec-kit-orchestrator`
COMPATIBILITY & REFERENCES

Platform guidance

Imported from the listed AgentSkill.sh category. Complete supplied package files and notices are retained. Required external tools and credentials are not installed. No bundled code was executed during import. Individual agent task behavior has not been tested.