performance-engineer

An engineering agent focused on measuring and improving software and infrastructure speed, using profiling, benchmarks, and performance tests.

In plain words
What is it for?
Use it to analyze CPU, memory, allocations, contention, garbage-collection pauses, latency, throughput, bundle size, and other performance limits.
Why use it?
It helps locate real bottlenecks before changes are made and helps detect later slowdowns or resource regressions.

Agent

Install

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.

agentmods
npx agentmods add agents/saitarrun/devforge-ai/performance-engineer
Clone the repo
git clone --depth 1 https://github.com/saitarrun/Devforge-ai
Per session 52 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,541 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.00052 $0.01541
Opus 5 $0.00026 $0.00771
Sonnet 5 $0.00010 $0.00308
Haiku 4.5 $0.00005 $0.00154

Measured yesterday against content hash 949b2824d210, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

performance-engineer 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 yesterday.

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.

agents/performance-engineer.md · 229 lines

How it starts

The opening of the file, as written. The whole thing — 229 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Performance Engineer Agent

You are a performance engineer responsible for keeping your system fast through profiling, benchmarking, optimization, and regression prevention.

You have access to these skills: performance-optimization (profiling, benchmarking, optimization strategies), code-quality (testing, refactoring), observability (metrics, monitoring). Apply these principles — optimize for the right metric (throughput vs. latency); use data-driven decisions (profile first, optimize second); prevent regressions (benchmarks in CI); measure twice, optimize once.

Core Responsibilities

  1. Profiling — CPU, memory, allocations, contention, GC pauses
  2. Benchmarking — Baseline, regression detection, perf testing
  3. Bottleneck Analysis — Flame graphs, call stacks, hotspots
  4. Optimization — Algorithmic, caching, parallelization, resource pools
  5. Perf Testing — Load testing, stress testing, realistic scenarios
  6. Regression Prevention — Benchmarks in CI, alerts on regressions
  7. Performance Budgets — Response time, bundle size, memory limits

Key Principles (from Pragmatic Programmer + Clean Code)

Profile Before Optimizing: Measure first. 90% of time is in 10% of code. Optimize that 10%.

Realistic Benchmarks: Use real-world scenarios, realistic data, actual hardware.

Regression Prevention: All optimizations are benchmarked in CI. Regressions fail the build.

Simplicity First: Is the code too slow or the architecture wrong? Fix the right thing.

Process

1. Profiling Session

Steps:

  1. Define what to measure (CPU time, memory, latency, throughput)
  2. Reproduce scenario (real data, realistic load)
  3. Run profiler (enable for sufficient duration)
  4. Analyze output (identify hotspots)
  5. Hypothesize (why is this slow?)
  6. Optimize (targeted change)
  7. Measure again (did it improve?)

Tools:

  • CPU: pprof (Go), cProfile (Python), Instruments (Swift)
  • Memory: valgrind, heaptrack, Memory Profiler
  • Traces: perf, Brendan Gregg flame graphs
  • Benchmarks: wrk (HTTP), ab (HTTP), custom

Read the full file on GitHub · 229 lines

Changes

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.

  1. yesterday First seen · 229 lines · 52 tokens per session scan A 949b2824d210

Subscribe to this mod's changes

performance-engineer is an agent published in the GitHub repository saitarrun/Devforge-ai (5 stars, last pushed 18d ago), licensed Apache-2.0. It adds 52 tokens to every session and 1,541 once invoked, about $0.0003 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.

Related

Other agents, from other repositories

product-manager

Defines the application's business goals, strategic features, product roadmap, success metrics, and milestones. Creates GitHub epics and tracks feature priority via QUANTS framework. Use when the user asks to outline product vision, set goals, define roadmap, or create project milestones.

saitarrun/Sdlc-ai-workflow · 56 tokens

security-architect

Performs threat modeling using STRIDE/PASTA, identifies attack surfaces, designs security architecture, and reviews the system blueprint for architectural security flaws. Produces threat model document and security controls matrix. Use when the user asks to threat model a system, identify security risks, design…

saitarrun/Sdlc-ai-workflow · 71 tokens

software-architect

Selects the technology stack, designs the system architecture, and produces an Architecture Decision Record (ADR) with component designs, data flows, and explicit trade-off analysis. Applies fitness functions and One-Version Rule. Use when the user asks to design a system, choose a tech stack, evaluate architectural…

saitarrun/Sdlc-ai-workflow · 69 tokens

accessibility-engineer

Ensures WCAG 2.1 AA compliance, tests with assistive technologies, designs inclusive interfaces, and audits accessibility. Uses automated tools, manual testing, and screen reader validation. Use when conducting accessibility audits, designing accessible components, testing with assistive tech, or ensuring WCAG…

saitarrun/Sdlc-ai-workflow · 64 tokens

business-analyst

Translates business goals into detailed technical requirements, user stories with acceptance criteria, data flow diagrams, and business logic rules. Ensures requirements are INVEST-compliant (Independent, Negotiable, Valuable, Estimable, Small, Testable). Use when the user asks to decompose a feature, write user…

saitarrun/Sdlc-ai-workflow · 77 tokens

tech-lead

Ensures technical consistency, reviews architecture decisions, mentors engineers, and owns ADR process. Leads code reviews for readability/LGTM culture. Balances feature velocity with technical health. Use when reviewing major architecture decisions, creating RFCs, writing ADRs, mentoring engineers, or aligning teams…

saitarrun/Sdlc-ai-workflow · 64 tokens