tpu_performance_autoresearch_wiki: Agent for Claude Code

.claude/agents/profile-analyzer.md

profile-analyzer is an agent for Claude Code from vlasenkoalexey/tpu_performance_autoresearch_wiki. It costs 0 tokens per session (6,170 once invoked), scanned A, original, MIT.

An analysis worker for completed machine-learning kernel experiments. It examines an xprof performance trace and an HLO compiler dump, then returns report sections for the experiment page.

In plain words
What is it for?
It is used to attribute time to operations, inspect memory behavior, review compiled modules and fusions, and report regressions or other findings from the trace and dump.
Why use it?
It makes profiling and compiler inspection a required part of the experiment record instead of relying only on headline timings or skipping analysis.

Agent for Claude Code

Written for Claude Code: installed under .claude/. Also seen: model in frontmatter.

This is vlasenkoalexey/tpu_performance_autoresearch_wiki's own configuration. It tells Claude Code how to work on tpu_performance_autoresearch_wiki itself, so it is not a mod to install elsewhere. Copy it as a starting point and replace the rules that are about this project. Everything tpu_performance_autoresearch_wiki configures →

Reuse

Borrowing it

Nothing to install: this file belongs to vlasenkoalexey/tpu_performance_autoresearch_wiki. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file
curl -O https://raw.githubusercontent.com/vlasenkoalexey/tpu_performance_autoresearch_wiki/main/.claude/agents/profile-analyzer.md
Clone the repo
git clone --depth 1 https://github.com/vlasenkoalexey/tpu_performance_autoresearch_wiki

Made for: Claude Code.

Wrote 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.

agentmods badge for profile-analyzer

README.md
[![agentmods](https://agentmods.dev/badge/agents/vlasenkoalexey/tpu_performance_autoresearch_wiki/profile-analyzer/github.svg)](https://agentmods.dev/agents/vlasenkoalexey/tpu_performance_autoresearch_wiki/profile-analyzer)
Your own site
<a href="https://agentmods.dev/agents/vlasenkoalexey/tpu_performance_autoresearch_wiki/profile-analyzer"><img src="https://agentmods.dev/badge/agents/vlasenkoalexey/tpu_performance_autoresearch_wiki/profile-analyzer/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for profile-analyzer

Your own site · 80×15
<a href="https://agentmods.dev/agents/vlasenkoalexey/tpu_performance_autoresearch_wiki/profile-analyzer"><img src="https://agentmods.dev/badge/agents/vlasenkoalexey/tpu_performance_autoresearch_wiki/profile-analyzer.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 0 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 6,170 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
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 5.1 $0.00000 $0.06170
Opus 5 $0.00000 $0.03085
Sonnet 5 $0.00000 $0.01234
Haiku 4.5 $0.00000 $0.00617

Measured 10d ago against content hash 06f2598105e2, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-10, from the pricing page.

Security

Grade A, and why

profile-analyzer 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 10d 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.

.claude/agents/profile-analyzer.md · 271 lines

How it starts

The opening of the file, as written. The whole thing — 271 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are a profile + HLO analyzer for the autoresearch loop. Your job is narrow and bounded: given one completed experiment's GCS paths, produce a structured ## Profile + ## HLO Dump markdown payload that the master agent slots directly into the experiment page.

The master agent owns the experiment-page filing, hypothesis selection, and verdict assignment. You ONLY analyze. You do not write wiki pages, do not propose next hypotheses, do not update the model page.

Program.md step 8 says profile + HLO analysis is MANDATORY — NON-NEGOTIABLE. You exist because that step was systematically skipped before. Do not return without attempting both phases.

What the master gives you

Field Example Required?
exp_slug <model>-jax-v391m-tokamax-ce YES
index_path /mnt/disks/persist/.../wiki/profile-analyzer-index.md (ABSOLUTE) YES
experiment_page_path /mnt/disks/persist/.../wiki/experiments/<model>_autoresearch_optimization/jax/2026-05-29-v391m-tokamax-ce.md (ABSOLUTE) YES
xprof_path gs://tpu-pytorch-alekseyv-us-central2/autoresearch/<slug>/xprof YES
hlo_path gs://tpu-pytorch-alekseyv-us-central2/autoresearch/<slug>/hlo YES
variant 24B/v5p-32 YES
current_frontier_mfu 21.61% (snapshot at dispatch time — may be 1 iteration stale if another track landed a verdict in parallel; treat as context, not authoritative) YES
focus "data_formatting grew 7pp — why?" (free-form question) optional (Mode B drilldown)

If exp_slug, index_path, experiment_page_path, xprof_path, or hlo_path is missing, refuse and ask the master to provide it. Do not improvise paths — relative paths and convention-derived subpaths break silently. Absolute paths are the contract.

Cross-experiment diffing is NOT in your scope. The prior_exp_slug field is intentionally absent — master handles cross-experiment comparison at verdict-assignment time, and any explicit cross-experiment investigation goes through Mode B (focus="diff:vX vs vY").

Read the full file on GitHub · 271 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. 10d ago First seen · 271 lines · 0 tokens per session scan A 06f2598105e2

Subscribe to this mod's changes

profile-analyzer is an agent published in the GitHub repository vlasenkoalexey/tpu_performance_autoresearch_wiki (55 stars, last pushed 6d ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 6,170 tokens. 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 2026-08-30.

Related

Other agents, from other repositories

pixel-art-animation-reviewer

Independent reviewer of pixel-art ANIMATION quality (loop seamlessness, motion physics, multi-component motion, frame timing, period selection, particle determinism). One of four specialized review roles in the pixel-art-quality-board orchestrator. Use when the user asks to "check animation timing", "verify loop…

AnastasiyaW/codex-claude-code-config · 140 tokens

community-ecologist

Reasons from Vellend's four processes and Chesson stabilizing/equalizing coexistence through PERMANOVA/betadisper, betapart turnover–nestedness, Gotelli SIM9/C-score null models, and vegan/entropart/picante pipelines while treating compositional closure, dispersion heterogeneity, and pseudoreplicated quadrats as…

K-Dense-AI/scientific-agents · 81 tokens

algorithms-researcher

Reasons from separating problem, model, and cost model (comparison, word-RAM, arithmetic, online) through exchange/matroid greedy proofs, subproblem-DAG dynamic programming, max-flow min-cut and Goemans–Williamson primal-dual rounding, Karp–Rabin fingerprinting, competitive ratio and Yao's principle, PTAS/FPTAS…

K-Dense-AI/scientific-agents · 163 tokens

astrochemist

Reasons from gas-grain reaction networks, H₂ ortho/para and CR ionization rates through KIDA/kida.uva.2024, CDMS/JPL/Splatalogue line lists, Nautilus/UCLCHEM gas-grain models, ALMA/JWST/LIDA ice–gas linkage, XCLASS LTE fitting, and line-blending discrimination—not generic chemistry.

K-Dense-AI/scientific-agents · 81 tokens

building-science-engineer

Reasons from coupled heat-air-moisture transport through ASHRAE 160 moisture-design analysis, WUFI transient simulation, ACH50 leakage mapping, and ISO 10211 psi-values while treating exfiltration condensation, reservoir claddings, and mold-index sensitivity as first-class failure modes.

K-Dense-AI/scientific-agents · 66 tokens

aerodynamicist

Reasons from circulation, Cp distributions, and boundary-layer physics through Re/Mach similitude, NACA airfoil polars, stall classification, wind-tunnel blockage/wall corrections, and SA/SST/LES external-aero CFD—not generic mechanical engineering.

K-Dense-AI/scientific-agents · 56 tokens