sw-direct

name	sw:direct
description	Design Rider-focused interventions -- bright spot replication, critical moves, and destination postcard. Reads the diagnosis and produces a plan to give the Rider clarity and direction. Use after /sw:diagnose, or for focused revision of Rider interventions.
argument-hint	[optional: focus area -- bright-spots, critical-moves, or destination]

<direct_context> #$ARGUMENTS </direct_context>

Direct the Rider

Design interventions for the rational mind. The Rider needs clarity, specific behaviors, and a vivid destination.

When to Use

• After /sw:diagnose when you want to focus on Rider interventions specifically
• Revising or deepening Rider interventions after an /sw:intervene pass
• When the diagnosis ranks Rider as the primary sticking point
• "The team doesn't know what to do differently"

Before Starting

Check for diagnosis

Read docs/switch/current-diagnosis.md. Hard redirect if missing (see references/shared-enforcement-prompts.md).

Check Rider ranking

Read the sticking point section. If Rider is "not a major factor," warn (see references/shared-enforcement-prompts.md -- Sticking Point Mismatch). Do not block.

Check for prior Rider interventions

Search docs/switch/interventions/ for files with type: intervention-rider. Surface relevant ones.

Check for missing dimensions

If Elephant and Path intervention files don't exist, note:

"You have Rider interventions but no Elephant or Path interventions yet. The Switch framework says all three must be addressed. Consider running /sw:intervene for full coverage, or /sw:motivate and /sw:shape after this."

Accept context

If arguments specify a focus area (bright-spots, critical-moves, destination), jump to that step. Otherwise, work through all three.

The Sequence

Work through all 3 interventions in order. Use references/enforcement-prompts.md for skill-specific redirects and references/shared-enforcement-prompts.md for cross-skill rules.

---

Step 1: Find the Bright Spots (Extract Replicable Practices)

"The diagnosis identified these bright spots: [from diagnosis Step 2]. Now let's extract the replicable practice. What's the specific, teachable behavior from the bright spot that others could adopt?"

Note: The diagnosis identified where bright spots exist. This step extracts the replicable practice -- the specific behavior that can be transplanted.

What you're looking for:

• A specific practice extracted from the bright spot (not "be more like them")
• Why it works in the bright spot context
• How to transplant it -- what needs to change for others to adopt it
• Potential barriers to adoption (preview for Elephant/Path interventions)

Enforcement triggers: See references/enforcement-prompts.md Step 1.

Use AskUserQuestion to confirm the replicable practice before moving on.

---

Step 2: Script the Critical Moves

"What are the 2-3 specific behaviors that, if people adopted them, would produce the change? Not principles -- behaviors. Not 'be more collaborative' -- 'start every standup by naming one blocker you need help with.'"

What you're looking for:

• 2-3 specific, observable behaviors (not values, not principles)
• Each behavior described as a trigger-action pair: "When [situation], do [this]"
• What the old behavior is that it replaces
• Why each behavior matters (connects to the desired change)

Enforcement triggers: See references/enforcement-prompts.md Step 2.

---

Step 3: Point to the Destination (Destination Postcard)

"Paint a vivid picture of what success looks like. Not metrics -- a scene. What does a typical day look like once this change has happened? Make it concrete enough that people can see themselves in it."

What you're looking for:

• A vivid, sensory description (not KPIs)
• Specific enough that people can picture themselves there
• Emotionally appealing (intentionally crosses into Elephant territory)
• A "destination postcard" someone could hold in their mind

Enforcement triggers: See references/enforcement-prompts.md Step 3.

---

Assembling the Rider Intervention Plan

After all 3 steps:

1. Read the template from references/rider-intervention-template.md
2. Fill in all sections
3. Check alignment with diagnosis: does each intervention address a specific Rider problem identified in the diagnosis?
4. Present for review

"Here's your Rider intervention plan. Do the critical moves feel actionable? Does the destination postcard make you want to get there?"

Saving

• Write to docs/switch/interventions/rider-{slug}-{date}.md
• Set frontmatter: type, tags, diagnosis reference, diagnosis date, status (designed), confidence (ask user), created date, source

After Saving

Use AskUserQuestion:

Question: "Rider intervention plan saved. What next?"

Options:

1. Run /sw:motivate -- The Rider knows where to go. Now motivate the Elephant.
2. Run /sw:shape -- Shape the Path to make the critical moves easy.
3. Run /sw:compound -- Save what you learned.
4. Revise the diagnosis -- Run /sw:diagnose if the interventions revealed gaps.
5. Done for now.

Important Rules

• Bright spot replication builds on the diagnosis. Don't re-identify bright spots -- extract practices from the ones already found.
• Critical moves must be observable. If you can't watch someone and confirm "they did it," it's too vague.
• Maximum 3 critical moves. The Rider gets overwhelmed by options.
• The destination postcard is a scene, not metrics. KPIs are for the Path; the postcard is for the Rider and Elephant.

Pipeline Mode

When invoked with disable-model-invocation context:

• Skip all AskUserQuestion prompts
• Use diagnosis bright spots and Rider assessment as input
• Make reasonable assumptions, flag them in output
• Write output file without waiting for confirmation