OpenParable
Skip to content
← Back to skills

Plain Optimize

Testing & Quality
plain-optimize

Plain Optimize is an Agent Skill for testing & quality published in the dropseed/plain repository on GitHub: a SKILL.md instruction file an AI agent can load to do this task.

Captures and analyzes performance traces to identify slow queries and N+1 problems. Use when a page is slow, there are too many queries, or the user asks about performance.

View on GitHub ↗
QUICK START

How to use this skill

Bring this guide into your coding agent with a prompt tailored to the tool you use.

  1. Open your project in Codex.
  2. Copy and paste the prompt below.
  3. Review files and risks before approving any action.
Prompt to paste
Inspect the Agent Skill at https://github.com/dropseed/plain/blob/HEAD/.claude/skills/plain-optimize/SKILL.md and all supporting files. Report provenance, permissions, scripts, dependencies, and security risks. Propose an exact installation into .agents/skills/plain-optimize, but wait for my approval before writing files or running scripts.

Copying this prompt does not install or run the skill. Review third-party files before use.

Performance Optimization Workflow

1. Capture and Analyze

Make a request and get structured JSON — response metadata plus a full trace analysis:

uv run plain request /path --json
uv run plain request /path --json --user 1
uv run plain request /path --json --method POST --data '{"key": "value"}'

The --user flag accepts a user ID or email.

The trace object has two parts — analysis (derived) and spans (raw):

  • analysis.query_count / analysis.duplicate_query_count — query summary
  • analysis.duration_ms — total trace duration
  • analysis.issues — pre-detected problems (N+1 queries, exceptions)
  • analysis.queries — each unique query with count, total duration, and source locations
  • spans — raw OpenTelemetry spans, a flat list (parent_span_id gives the structure)

2. Identify Bottlenecks

Check analysis.issues first — duplicate queries are flagged automatically with source locations. Then review:

  • N+1 queries (duplicate queries with count > 1)
  • Slow database queries (high total_duration_ms)
  • Missing indexes
  • Unnecessary work in hot paths

3. Apply Fixes

  • Add select_related() / prefetch_related() for N+1
  • Add database indexes for slow queries
  • Cache expensive computations

4. Verify Improvement

Re-run uv run plain request /path --json and compare analysis.query_count, analysis.duplicate_query_count, and analysis.duration_ms.