Stop repeating yourself to AI. Save what worked. Reuse it forever.

Experimental spec: v1.0-exp

AEP saves successful AI workflows in your repo so future runs start aligned instead of from zero.

Every AI session starts from scratch.

You keep repeating:

• "do not redesign the layout"
• "reuse the existing CSS"
• "keep this practical, not over-engineered"

...even when you already solved this exact task last week.

AEP fixes that.

Save what worked once. Reuse it forever across sessions, tasks, and agents.

Task -> Successful outcome -> Save AEP -> Future tasks start aligned

AEP makes AI behave like it already learned your preferences.

Without AEP, you usually say:

"Convert this HTML landing page to Next.js. Do not redesign. Keep CSS. Avoid premature componentization."

With AEP, you say:

"Use the AEP skill and apply packs for HTML -> Next.js before you start."

And the agent starts with your intent, constraints, workflow, and checks already loaded.

Tell your agent:

• "Use the AEP skill and save this collaboration as html-to-nextjs-migration."
• "Use the AEP skill and apply relevant packs before you start."
• "Use the AEP skill and explain which AEPs are active right now."

AEP packs live in .agent/aep/ as repo-local JSON.

Install page: smithery.ai/skills/Robi-Labs/aep

Install with Skills CLI:

npx @smithery/cli@latest skill add Robi-Labs/aep

You can also install directly from the Smithery skill page into your agent of choice.

• Convert HTML -> Next.js without redesign
• Write consistent landing page copy
• Generate UI that matches your style
• Reuse backend refactor patterns
• Standardize team workflows across projects

• Less repetition
• Better first drafts
• Consistency across tools
• Lower token cost over time

flowchart LR
  U[Task] --> A[Apply AEP]
  A --> M[Match Packs]
  M --> R[Constraints + Workflow]
  R --> W[Execute Task]
  W --> S[Save AEP]
  S --> P[Next tasks improve]

.agent/
  aep/
    index.json
    project.aep.json
    user.aep.json
    tasks/

Time ->
          ________
Quality  /
        /
       /
      /
_____/________________

With AEP: improves over time
Without AEP: resets every session

• SKILL.md
  v1.0-exp behavior for an AEP-aware agent
• references/schema.v1.md
  full pack and index schema for version: "1.0-exp"
• references/integration-notes.v1.md
  integration guidance for matching, scoring, and compatibility
• references/prompt-patterns.v1.md
  prompt patterns for save/apply/promote/inspect flows
• assets/templates/
  template files for project/user/task/index

• Better matching:
  • applies_to.languages, frameworks, paths, domains
  • strength score per pack
• Lightweight analytics:
  • metrics.times_applied, first_used_at, last_used_at, avg_turns_saved
• Pack evolution:
  • history events
  • optional merge_suggestions

Everything stays repo-local, file-only, and agent-agnostic.

1. Create:

.agent/
  AGENTS.md
  aep/

2. Copy templates:

• aep/assets/templates/project.aep.json -> .agent/aep/project.aep.json
• aep/assets/templates/user.aep.json -> .agent/aep/user.aep.json
• aep/assets/templates/index.json -> .agent/aep/index.json
• optionally:
  • aep/assets/templates/task.generic.aep.json -> .agent/aep/tasks/<task>.aep.json

3. Agent behavior:

• read .agent/AGENTS.md
• read .agent/aep/*.json
• follow aep/SKILL.md

• aep/SKILL.md
• aep/references/schema.v1.md
• aep/references/integration-notes.v1.md
• aep/references/prompt-patterns.v1.md