Skip to content

demo-assets.md

Latest commit

 

History

History
74 lines (61 loc) · 2.6 KB

demo-assets.md

File metadata and controls

74 lines (61 loc) · 2.6 KB

Demo Assets Plan

This document tracks demo assets that should be recorded with real commands and reproducible outputs. Demos should be short, terminal-first, and technical.

Standards

  • Use real commands and visible output.
  • Keep terminal text readable at 1080p.
  • Include command transcripts below GIFs or videos.
  • Do not hide failed commands; rerun and explain if needed.
  • Prefer asciinema or terminal recordings that can be regenerated.

Demo 1: First Run In A Fresh Repository

  • Goal: show that AMS can be initialized quickly.
  • Scenario: run AMS inside a small TypeScript or Node repository.
  • Commands:
    • npx @ravbyte/agent-memory-system@latest init
    • tree memory or Get-ChildItem memory
    • open memory/context-index.json
  • Expected output:
    • generated memory/ files
    • AGENTS.md
    • context index with topic mappings
  • Ideal duration: 35-45 seconds.

Demo 2: Graph Blast-Radius Query

  • Goal: show why graph artifacts matter before shared-interface edits.
  • Scenario: inspect downstream impact of src/types.ts.
  • Commands:
    • agent-memory graph build
    • agent-memory graph query --file src/types.ts
    • optionally agent-memory graph diff
  • Expected output:
    • affected files
    • dependency direction
    • graph summary
  • Ideal duration: 45-60 seconds.

Demo 3: Worklog Checkpoint

  • Goal: show agent progress tracking as plain repo artifacts.
  • Scenario: record a checkpoint after a small change.
  • Commands:
    • agent-memory worklog checkpoint --agent codex --message "updated scanner docs" --files README.md --commands "npm test"
    • Get-Content memory/agent-worklog.jsonl -Tail 1
  • Expected output:
    • JSONL checkpoint with agent, message, files, and commands.
  • Ideal duration: 35-45 seconds.

Demo 4: Cross-Agent Handoff

  • Goal: show how another agent/session resumes from state.
  • Scenario: agent A stops after implementation; agent B reads handoff and continues.
  • Commands:
    • agent-memory worklog handoff --agent codex --message "tests pass; docs need review" --next "review README and update examples"
    • Get-Content memory/agent-handoff.md
  • Expected output:
    • human-readable current state, recent events, next steps, and files mentioned.
  • Ideal duration: 60 seconds.

Demo 5: Benchmark Run

  • Goal: explain benchmark design without overstating results.
  • Scenario: run one fixture task in baseline and AMS mode.
  • Commands:
    • npx tsx benchmarks/scripts/run.ts --repo . --category understanding --mode both
    • npx tsx benchmarks/scripts/report.ts
  • Expected output:
    • run JSON
    • markdown report
    • metrics table with limitations
  • Ideal duration: 75-90 seconds.