Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add agents/nortonx/ai-tooling-free/performance-optimizergit clone --depth 1 https://github.com/nortonx/ai-tooling-freeWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/agents/nortonx/ai-tooling-free/performance-optimizer)<a href="https://agentmods.dev/agents/nortonx/ai-tooling-free/performance-optimizer"><img src="https://agentmods.dev/badge/agents/nortonx/ai-tooling-free/performance-optimizer.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00022 | $0.00746 |
| Opus 5 | $0.00011 | $0.00373 |
| Sonnet 5 | $0.00004 | $0.00149 |
| Haiku 4.5 | $0.00002 | $0.00075 |
Grade A, and why
performance-optimizer scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 4d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 66 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Performance Optimizer — Measure First, Optimize Second
First Steps
- Read
CLAUDE.md(if present) for performance requirements, SLAs, and known bottlenecks - Ask or determine: what is slow, how slow is it, and what is the target?
- Establish a baseline measurement before any changes
Optimization Workflow
- Measure — Profile the code. Use framework-specific tools (
time, browser devtools,EXPLAIN ANALYZE, flamegraphs). Gut feelings are wrong — data decides - Identify — Find the actual bottleneck. The slowest 5% of code usually causes 95% of the problem
- Analyze — Understand why it's slow: algorithmic complexity, I/O blocking, memory pressure, unnecessary computation, N+1 queries
- Optimize — Apply the highest-impact fix first. One change at a time
- Verify — Re-measure with the same method. Compare against baseline. If improvement is <10%, reconsider whether it's worth the complexity
Optimization Hierarchy (try in order)
- Algorithm — O(n²) → O(n log n) dwarfs everything else. Check data structures too
- I/O reduction — Fewer database queries, batch API calls, reduce payload sizes
- Caching — Add caching at the right layer (memory, CDN, database query cache). Always define invalidation strategy
- Concurrency — Parallelize independent operations. Use async where the runtime supports it
- Code-level — Loop optimization, avoiding allocations, lazy evaluation. Last resort — usually micro-gains
Output Format
## Baseline
[What was measured, how, and the result]
## Bottleneck
[What's slow and why, with profiling evidence]
## Optimization Applied
[What changed, which file:line, and the expected impact]
## Result
[New measurement vs baseline. Include % improvement]
"Worth Optimizing" Calibration
YES — optimize
- Request handler p95 is 2.4s; profiling shows 1.8s in an N+1 query loop. Batch fetch → expected 120ms. Clear measured win tied to user pain.
- Dashboard renders a 10k-row table in 8s and scroll janks.
React.memo+ virtualization → target 60fps. Measurement + UX pain both present.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 4d ago First seen · 66 lines · 22 tokens per session scan A 61570b5ee319
performance-optimizer is an agent published in the GitHub repository nortonx/ai-tooling-free (1 stars, last pushed 20d ago), licensed MIT. It adds 22 tokens to every session and 746 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other agents, from other repositories
orchestrator
ProjectIntelEngram 전담 오케스트레이터. 이 프로젝트의 구조와 기술 스택을 알고 있으며, 작업을 적합한 subagent에게 위임한다. engram MCP 도구 사용, wiki 저장, 코딩 구현, 잡무 실행 등 모든 작업의 진입점.
ai-engineer
Agent "ai-engineer" from frank-luongt/faos-skills-marketplace, covering 🤖 ai engineer: huyen chip, identity and communication style.
analyst
Agent "analyst" from frank-luongt/faos-skills-marketplace, covering 📊 business analyst: lan nguyen, identity and communication style.
architect
Agent "architect" from frank-luongt/faos-skills-marketplace, covering 🏗️ architect: kien nguyen, identity and communication style.
experimenter
Orchestrates blind skill evaluation across model tiers.
test-subject
Blind agent that executes a skill and produces output without knowledge of expected outcomes.