skills.rest
spideynolove
spideynoloveCOMMUNITY
Explore all skills by this author →
Website𝕏@spideynoloveGitHub

rlm-subagent

Dispatches parallel subagents for implementation, code review, and testing with sequential fallback.

2|3|Updated Mar 13, 2026
Search Tags:#testing#tdd#code review#workflow orchestration#parallel execution#subagent#fallback
One-click install
npx skills add https://github.com/spideynolove/claude-dotfiles --skill rlm-subagent-spideynolove
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: rlm-subagent
Source: https://github.com/spideynolove/claude-dotfiles/tree/main/.agents-global/skills/rlm-workflow/skills/rlm-subagent
Command: npx skills add https://github.com/spideynolove/claude-dotfiles --skill rlm-subagent-spideynolove

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Coordinating multi-phase implementation work across independent sub-phases is slow and error-prone when done sequentially in a single context, and teams lack a structured way to parallelize work while preserving review quality and TDD discipline. ## Core Features & Use Cases - Parallel Subagent Dispatch: Spawns implementer subagents per independent sub-phase (SP1, SP2, SP3) and aggregates results with two-stage spec and code-quality review. - Automatic Sequential Fallback: Detects subagent capability at runtime and falls back to sequential execution with extended self-review when subagents are unavailable. - Phase 3.5 Code Review Loop: Dispatches a dedicated code-reviewer subagent with severity classification (Critical/Important/Minor) and re-review cycles until approval. - Use Case: A feature touching API, UI, and database layers is split into three independent sub-phases, each implemented by a separate subagent in parallel, reviewed, then validated with parallel unit, integration, and E2E test subagents. ## Quick Start Ask the agent to parallelize Phase 3 implementation by dispatching an implementer subagent for each independent sub-phase in the locked TO-BE plan.

Frequently Asked Questions about rlm-subagent

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I parallelize implementation across multiple sub-phases?▼

Extract independent sub-phases from the locked Phase 2 TO-BE plan and dispatch one implementer subagent per sub-phase in parallel. Each subagent receives a self-contained prompt with the full SP text, BASE_SHA, and relevant context, then follows TDD discipline before reporting completion.

How to run a code review subagent before testing?▼

Dispatch a code-reviewer subagent in Phase 3.5 with the git range (BASE_SHA..HEAD_SHA), the Phase 2 plan, and the implementation summary. It classifies issues as Critical, Important, or Minor and returns a verdict of Approved or Changes Required, looping until approval.

What happens when subagents are not available on my platform?▼

The skill detects subagent capability at the start of Phase 3 and falls back to sequential mode automatically. The main agent then executes each sub-phase in order with an extended self-review checklist and integration testing between sub-phases.

When should I use sequential mode instead of parallel subagents?▼

Use sequential mode when subagents are unavailable, when there is only a single sub-phase, or when sub-phases have dependencies on each other's output. Parallel dispatch only benefits workflows with three or more independent sub-phases.

Why do integration tests fail after parallel sub-phase implementation?▼

Failures usually come from file conflicts between sub-phases, shared mutable state, or commit ordering issues. Review the commit history, verify sub-phases touched isolated files, and check that each implementer subagent received a correct BASE_SHA.

Related Skills

Explore More Software Engineering Skills→
SSiertsemaSSiertsema

story-splitting

Decompose large user stories into INVEST-compliant sub-stories with SPIDR patterns.

Community
Advanced
p-o-ke-naep-o-ke-nae

github-flow

Guide GitHub Flow branching, commits, testing, pull requests, and squash merges.

Community
Intermediate
doubleuuserdoubleuuser
67

rlm-workflow

Orchestrate phase-gated RLM runs with SHA-256 LockHash verification and artifact locking.

Community
Advanced
try-workstry-works
4

rlm-subagent

Coordinate parallel RLM sub-phase execution with sequential fallback and two-stage reviews.

Community
Advanced