# qtr2-control audit: blocked during required setup

Credit: Taylor (gtmcknight), author of the solved policy identified by the assignment. No optimization patch was made.

```json
{"blocked":{"step":"sh scratch/setup.sh: initial mkdir","error":"Required setup script exited 1: mkdir: /Users/longqiaofeng/Library/Application: Operation not permitted (three occurrences). The script expands unquoted $R under a workspace path containing spaces."}}
```

This is an observed local setup failure, not a request for an owner address, key, or other requester input. The assignment explicitly requires returning blocked={step,error} on setup failure; setup was not repaired or retried. The host reports Darwin/arm64; the requested engine setup targets Linux x86_64. Compatibility beyond the initial failure was not tested.

## Work performed and evidence

- Cloned the challenge repository into `test/scratch/amm-challenge`, inside the workspace, and checked out the requested revision `6b6921f8d463314aa7c96234e1be25a25245f704`. Checkout completed with exit 0 and reported `HEAD is now at 6b6921f`.
- Saved the supplied setup text as `scratch/setup.sh` in that checkout and executed it with `sh scratch/setup.sh`. The actual setup script was executed once. An earlier shell command used an incorrect relative destination, failed to create the script, and did not execute that script.
- Setup stopped at `set -eu;R=$PWD/scratch;mkdir -p $R/tmp $R/bin $R/solcx`. No environment file, uv download, wheel verification, Python environment, compiler installation, or strategy validation was reached by that script.
- Setup stdout/stderr was checkpointed at `test/scratch/amm-challenge/scratch/setup.log`. Scratch files are temporary and will not be submitted.
- The two artifact URLs were fetched into scratch, but their contents were not inspected or used to establish any policy or performance claim. Neither expected Solidity hash was verified.

The failure follows directly from shell word splitting of the supplied unquoted `$R` in this workspace. The error above is copied from the local command output. No external source passage is used as evidence for a completed audit; `sources.json` therefore contains an empty sources array.

## Requested audit status

| Deliverable/check | Observed status |
| --- | --- |
| P: reproduce Taylor gates with `python3 -I` | Not run; 0 completed |
| Verify qt1 Solidity SHA256 | Not performed |
| Verify qtr2 Solidity SHA256 | Not performed |
| Exact patch against R | No policy edits made; no extracted candidate or verified byte comparison available |
| Policy, tables, rounding | No changes made |
| Compile with challenge compiler / validate sources | Not run |
| F: seeds 70000..70049, first 2,000 callbacks each | 0 callbacks checked; entire range incomplete |
| Init and both fees, fresh deployments | Not checked |
| F maximum fee error / raw equality | Unavailable, not zero and not a pass |
| Runtime bytecode sizes | Unavailable |
| afterSwap peak / OOG count | Unavailable; limits not verified |
| Engine seeds 9800320..9800447 | No arms run; entire range incomplete for qt1, qtr2, candidate |
| Consumed engine seeds / reruns | None / none |
| Per-seed edges | Unavailable |
| Paired n, mean_diff, se, worst5 versus qt1 | n=0; all statistics unavailable |
| Raw gas_stats | Not collected |
| Replay and engine scripts | Not implemented after required setup failure |
| Timing | Setup runner reported 0.0 s elapsed (tool precision); no replica or engine timing exists |

No performance improvement, equivalence, gas compliance, bytecode compliance, or gate pass is claimed. This report is an incomplete audit, not a certified unchanged-control result. The two named output files are the only permanent deliverables; no contract or configuration was modified.