Borrowing it
Nothing to install: this file belongs to quangdang46/openproxy. 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/quangdang46/openproxy/main/.claude/skills/loop-test-fix/SKILL.mdgit clone --depth 1 https://github.com/quangdang46/openproxyWrote 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/quangdang46/openproxy/loop-test-fix)<a href="https://agentmods.dev/skills/quangdang46/openproxy/loop-test-fix"><img src="https://agentmods.dev/badge/skills/quangdang46/openproxy/loop-test-fix/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/quangdang46/openproxy/loop-test-fix"><img src="https://agentmods.dev/badge/skills/quangdang46/openproxy/loop-test-fix.svg" alt="Reviewed on agentmods" width="80" 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 5.1 | $0.00130 | $0.02350 |
| Opus 5.5 | $0.00052 | $0.00940 |
| Sonnet 5.5 | $0.00026 | $0.00470 |
| Haiku 4.5 | $0.00013 | $0.00235 |
Grade A, and why
loop-test-fix 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 1mo 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 — 135 lines — stays where its author put it; the contents beside it link to each section on GitHub.
openproxy loop-test-fix
Self-driving verification loop. No orchestration script — just follow these steps in order, each turn, using your own judgment for the "research → fix" part when something breaks. Keep looping until the exit condition is met or you're blocked and need the user.
Loop body
-
Build + install openproxy locally
cargo build --release(or the project's usual release build).- Install/link the freshly built binary so it's the one on PATH / used by codex config
(check
install.sh/update-openproxy.shfor the project's own install convention — prefer those over ad-hoc copying). - If the build fails: that's a bug — go to step 5 (research → fix) before continuing.
-
Set up the
opencodeprovider (free/no-auth) — NOTopencode-gooropencode-zen- openproxy has three distinct providers that all sound alike — do not conflate them:
opencode(aliasoc) — the free, no-auth one. This is the one to use for this loop.opencode-go— a separate provider, different base URL (https://opencode.ai/zen/go/v1).opencode-zen— requires a real API key (confirmed 401 without one) — never treat this as "free". Using the wrong one invalidates the whole test — a response fromopencode-go/opencode-zendoes not prove the free no-auth path (opencode) works.
- Before wiring anything up, verify
opencodestill has quota/isn't rate-limited (hit its/v1/modelsor do a tiny throwaway call). Do not fire the real smoke-test prompt blind — a burned-out free quota should be treated as "provider temporarily unavailable, retry later", not a failure of openproxy itself. - In openproxy's own config (provider setup / combo selection — see
AGENTS.mddashboard workflow: configure provider → customize models → create combo → select for opencode CLI), make sure a combo backed by theopencodeprovider is what's selected/exposed.
- openproxy has three distinct providers that all sound alike — do not conflate them:
-
Point Codex at openproxy — via openproxy's own CLI, not manual config.toml edits
- openproxy runs locally on port 4623 by default (
http://127.0.0.1:4623). - Use openproxy's built-in CLI integration to write Codex's
config.tomlfor you:openproxy tool apply codex --model "<combo/model id that resolves to the opencode provider>" \ --endpoint http://127.0.0.1:4623- Omit
--endpointif the server is already running on the default port (4623) — it defaults to the running server's URL. - Use
--dry-runfirst if you want to eyeball the JSON body before it POSTs. openproxy tool show codexafterwards confirms what got saved.
- Omit
- If unsure which model id resolves to the
opencodeprovider, checkopenproxy tool run provider-list/ the model catalog before runningapply— don't guess. - To undo/reset after testing:
openproxy tool revert codex.
- openproxy runs locally on port 4623 by default (
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.
- 1mo ago First seen · 135 lines · 130 tokens per session scan A 1c6095e60e6e
loop-test-fix is a skill published in the GitHub repository quangdang46/openproxy (29 stars, last pushed 5d ago), licensed MIT. It adds 130 tokens to every session and 2,350 once invoked, about $0.0005 per session on Opus 5.5. 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 2026-09-19.
Other skills, from other repositories
reproduce-bug
Reproduce a reported bug in googleapis/mcp-toolbox and decide whether it is real, delivering an evidence-backed verdict: confirmed, already fixed, misconfiguration, client-side, works as intended, not reproducible, or blocked. Use whenever a maintainer asks you to reproduce, verify, confirm, or investigate a bug…
verify-omnigent
Drive Omnigent the way a user does and prove a behavior with recorded evidence, using an isolated server, runner, and mock model plus a feature map of every user entry point. Load before reproducing a user-facing bug, before claiming a fix works, or when reviewing whether a change covered every surface (session vs.…
testing-blocks
Use this when you have made AEM Edge Delivery Services code changes to blocks, scripts, or styles and need to validate them before opening a pull request. Covers unit testing for utilities and logic, browser testing with Playwright, linting, and guidance on what to test and how.
reality-verification
This skill should be used when the user asks to "verify a fix", "reproduce failure", "diagnose issue", "check BEFORE/AFTER state", "VF task", "reality check", "check test quality", "mock-only tests", or needs guidance on verifying fixes by reproducing failures before and after implementation, or detecting mock-heavy…
webapp-qa
Perform scoped exploratory QA of a web application, reproduce defects, collect screenshots and console evidence, and deliver a prioritized report.
test-electron-app
Drive the real running PostHog Electron app (live tRPC, workspace-server, real data) over CDP with agent-browser. Connect to the running app on port 9222, test desktop changes against a local Django stack, snapshot the accessibility tree, inspect network requests, and screenshot only when explicitly asked. Use when…