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.
curl -O https://raw.githubusercontent.com/vlasenkoalexey/tpu_performance_autoresearch_wiki/main/.claude/skills/stop-experiment/SKILL.mdgit clone --depth 1 https://github.com/vlasenkoalexey/tpu_performance_autoresearch_wikiWrote 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/skills/vlasenkoalexey/tpu_performance_autoresearch_wiki/stop-experiment)<a href="https://agentmods.dev/skills/vlasenkoalexey/tpu_performance_autoresearch_wiki/stop-experiment"><img src="https://agentmods.dev/badge/skills/vlasenkoalexey/tpu_performance_autoresearch_wiki/stop-experiment/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.
<a href="https://agentmods.dev/skills/vlasenkoalexey/tpu_performance_autoresearch_wiki/stop-experiment"><img src="https://agentmods.dev/badge/skills/vlasenkoalexey/tpu_performance_autoresearch_wiki/stop-experiment.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
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 5.1 | $0.00070 | $0.04249 |
| Opus 5 | $0.00035 | $0.02124 |
| Sonnet 5 | $0.00014 | $0.00850 |
| Haiku 4.5 | $0.00007 | $0.00425 |
Grade A, and why
stop-experiment 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 11d 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 — 210 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are stopping the autoresearch loop cleanly. Follow this sequence — do not abbreviate.
Step 1 — Stop the loop
Step 1·0 — Final audit gate (BLOCKING — before the watcher is disarmed)
You are the agent the watcher supervises; you may not end your own supervision unaudited. Dispatch the process-auditor SYNC over the closing delta (the stop/at-ceiling claim, retrospective, and evidence table included). Then:
stop blocked/ FABRICATION finding → ABORT/stop-experiment. The watcher stays armed; apply the corrections and return to the loop (kernel family: the failing gate is the next K2 input). Exception: a user-ordered infrastructure stop making NO at-ceiling claim converts to a PAUSE (Stop rule's PAUSE ≠ STOP) instead of aborting — record the resume state and continue shutdown. A PAUSE claims no close, so the.stop-authorizedrequirement does NOT apply to it: disarm is permitted (the pause marker records the family OPEN; LINT's unauthorized-stop check keys on clean-close/at-ceiling markers only — without this carve-out a paused session could never disarm and the revive service would fight the operator's pause).- ALL-CLEAR (or only non-blocking findings, applied) → the auditor writes
<family>/pallas/.stop-authorized(its ALL-CLEAR line + cursor SHAs). VERIFY THE FILE EXISTS before proceeding —test -e <family>/pallas/.stop-authorized. No file = the audit did not complete clean = you may NOT proceed (wait for it or apply its findings; dispatching the audit is not the gate — its written authorization is). You may NEVER create this file yourself — that is fabrication-class and voids the close. Also record the auditor's report line verbatim; Step 6's marker cites it. A marker without a matching.stop-authorizedis an unauthorized stop (LINT voids + reopens). Proceed to the disarm.
Step 1·1 — Disarm the watcher
Only after Step 1·0 passes AND <family>/pallas/.stop-authorized exists on disk (or Step 1·0 resolved to a PAUSE — no close claimed), disarm the Step 9·0 process watcher — the persistent, self-rescheduling process-auditor task that /start-experiment armed at launch. Cancelling the watcher task at ANY other moment — including as "teardown hygiene" while an audit is still running — is a process violation that voids the close (the 2026-07-21 cohort's signature failure: three sessions cancelled their auditor mid-final-audit and closed unaudited). Its revive service would otherwise wake the session back up after this clean shutdown:
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.
- 11d ago First seen · 210 lines · 70 tokens per session scan A 851c4c9bebc2
stop-experiment is a skill published in the GitHub repository vlasenkoalexey/tpu_performance_autoresearch_wiki (55 stars, last pushed 7d ago), licensed MIT. It adds 70 tokens to every session and 4,249 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 2026-08-30.
Other skills, from other repositories
ml_inference_optimization
This document describes the Leeroopedia MCP tools available during the with-KB benchmark run. It is a standalone reference and is not fed to the agents automatically.
leeroopedia-mcp
Use Leeroopedia MCP to fetch grounded ML/AI best practices, build and review ML plans, debug failures, verify code/math correctness, and expand KB citations via getpage.
ml-research-lab
Machine-learning research loop for dataset curation, fine-tuning, evaluation, inference deployment, experiment tracking, and model explainability. Use when working on ML experiments, training data, model benchmarks, RunPod/GPU runs, classifier quality, vLLM/GGUF serving, SHAP-style model explanations, or…
experiment-audit
Use this skill for scientific and ML-research reasoning work — evaluating experimental claims, auditing training runs or ablations, checking whether a statistical claim holds up, assessing reproducibility, reconciling contradictory results, reviewing a paper's methodology or results section, writing reviewer-style…
midmem-research-tracker
Keep the repo's RESEARCH.md ledger current from the research ingestions already in the MidMem store: for every unevaluated weekly research ingestion, evaluate each paper's feasibility and applicable improvement for MidMem, and record it as Paper (cite + link) · Finding (the empirical claim that matters) · Decision…
sematyp
SemaTyP combines two data sources into a knowledge graph for drug discovery / repositioning.