Simple Task Workflow

Simple Task Workflow

Minimal path for straightforward tasks that don't require full orchestration complexity.

Overview

Not every task requires the full 6-stage workflow. Simple tasks can follow a streamlined path while still maintaining quality.

flowchart LR
    INPUT["Input"]
    EXECUTE["Execute"]
    VERIFY["Verify"]
    OUTPUT["Output"]

    INPUT --> EXECUTE --> VERIFY --> OUTPUT

When to Use Simple Workflow

flowchart TD
    START["New task arrives"]

    Q1{"Single, clear<br/>deliverable?"}
    Q2{"No research<br/>needed?"}
    Q3{"No delegation<br/>required?"}
    Q4{"Familiar<br/>domain?"}

    SIMPLE["→ Simple Workflow"]
    FULL["→ Full Workflow"]

    START --> Q1
    Q1 -->|"Yes"| Q2
    Q1 -->|"No"| FULL
    Q2 -->|"Yes"| Q3
    Q2 -->|"No"| FULL
    Q3 -->|"Yes"| Q4
    Q3 -->|"No"| FULL
    Q4 -->|"Yes"| SIMPLE
    Q4 -->|"No"| FULL

Simple Task Criteria:

• Single, well-defined deliverable
• No context gathering needed
• No sub-agent delegation
• Familiar technology/patterns
• Clear acceptance criteria

Simple Task Flow

flowchart TB
    subgraph INPUT["1. INPUT"]
        I1["Receive request"]
        I2["Confirm understanding"]
    end

    subgraph EXECUTE["2. EXECUTE"]
        E1["Implement solution"]
        E2["Use tools directly"]
    end

    subgraph VERIFY["3. VERIFY"]
        V1["Check deliverable"]
        V2["Run /verify if needed"]
    end

    subgraph OUTPUT["4. OUTPUT"]
        O1["Present result"]
        O2["Commit if appropriate"]
    end

    INPUT --> EXECUTE --> VERIFY --> OUTPUT

Detailed Sequence

sequenceDiagram
    participant U as User
    participant O as Orchestrator

    U->>O: Simple request
    O->>O: Confirm: Single deliverable?
    O->>O: Confirm: No research needed?
    O->>O: Confirm: Familiar domain?

    alt All Yes
        Note over O: Simple Workflow
        O->>O: Execute directly
        O->>O: Quick verify
        O->>U: Deliver result
    else Any No
        Note over O: Full Workflow
        O->>O: Use full 6-stage process
    end

Assets Used in Simple Workflow

Minimal Asset Path:

Input → [Direct Execution] → Quick Check → Output

Enhanced Asset Path:

Input → [Direct Execution] → /verify → git-commit-helper → Output

Example Simple Tasks

Example 1: Fix Typo

flowchart LR
    A["User: Fix typo<br/>in README"]
    B["Read file"]
    C["Edit file"]
    D["Present fix"]

    A --> B --> C --> D

No extra assets needed. Direct execution is sufficient.

Example 2: Add Function

flowchart LR
    A["User: Add helper<br/>function"]
    B["Read context"]
    C["Write function"]
    D["Quick verify"]
    E["Present code"]

    A --> B --> C --> D --> E

Optional: Use /verify before presenting if function is non-trivial.

Example 3: Update Config

flowchart LR
    A["User: Update config<br/>value"]
    B["Read config"]
    C["Edit value"]
    D["Validate format"]
    E["Present change"]

    A --> B --> C --> D --> E

No extra assets needed. Format validation is direct.

Escalation Triggers

Switch to full workflow if:

flowchart TD
    SIMPLE["Simple workflow<br/>in progress"]

    T1["Unexpected complexity<br/>discovered"]
    T2["Research becomes<br/>necessary"]
    T3["Multiple deliverables<br/>emerge"]
    T4["Delegation would<br/>help"]

    FULL["Switch to<br/>Full Workflow"]

    SIMPLE --> T1 --> FULL
    SIMPLE --> T2 --> FULL
    SIMPLE --> T3 --> FULL
    SIMPLE --> T4 --> FULL

Escalation Indicators:

• "I need to research this first"
• "This requires multiple steps"
• "An agent would handle this better"
• "I'm uncertain about the approach"

Quality Gates for Simple Tasks

Even simple tasks need minimal verification:

flowchart TD
    COMPLETE["Task appears<br/>complete"]

    G1{"Deliverable<br/>matches request?"}
    G2{"No obvious<br/>errors?"}
    G3{"Would I be<br/>satisfied as user?"}

    DONE["✅ Deliver"]
    REVIEW["⚠️ Review again"]

    COMPLETE --> G1
    G1 -->|"Yes"| G2
    G1 -->|"No"| REVIEW
    G2 -->|"Yes"| G3
    G2 -->|"No"| REVIEW
    G3 -->|"Yes"| DONE
    G3 -->|"No"| REVIEW

Simple vs Full Workflow Comparison

Decision Matrix

Common Patterns

Pattern: Quick Edit

1. Read target file
2. Make edit
3. Present change

Pattern: Add Small Component

1. Read existing code for context
2. Write new component
3. Quick verify syntax/patterns
4. Present result

Pattern: Configuration Change

1. Read current config
2. Make change
3. Validate format
4. Present updated config

Navigation

• Previous: Multi-Agent Orchestration
• Next: Investigation Workflow
• Back to: Index