Borrowing it
Nothing to install: this file belongs to ashishpatill/tell-proof. 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/ashishpatill/tell-proof/master/.cursor/skills/design-research-loop/SKILL.mdgit clone --depth 1 https://github.com/ashishpatill/tell-proofWrote 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/ashishpatill/tell-proof/design-research-loop)<a href="https://agentmods.dev/skills/ashishpatill/tell-proof/design-research-loop"><img src="https://agentmods.dev/badge/skills/ashishpatill/tell-proof/design-research-loop/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/ashishpatill/tell-proof/design-research-loop"><img src="https://agentmods.dev/badge/skills/ashishpatill/tell-proof/design-research-loop.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 | $0.00070 | $0.01777 |
| Opus 5.5 | $0.00028 | $0.00711 |
| Sonnet 5.5 | $0.00014 | $0.00355 |
| Haiku 4.5 | $0.00007 | $0.00178 |
Grade A, and why
design-research-loop 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 — 149 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Design research loop
The design engine is only allowed to get better in one way: by measuring pages that real buyers already paid for, turning those measurements into corridors, and then closing the distance between our output and those corridors. Opinions do not move the engine. Numbers do.
Run one loop per session. A loop is small, complete, and leaves the repository provably better than it found it.
Invariants
- Anonymised forever. The corpus of reference URLs lives in
research/corpus.local.json, which is git-ignored. Nothing committed — code, docs, comments, commit messages, test names — may contain a third-party person, studio, company, product, host, or URL. Committed research carriesref-0NNids and category buckets only. - Evidence before opinion. Any change to tokens, layout, motion, or copy defaults must cite a
measured band in
docs/10_DESIGN_EVIDENCE.md. "It looks better" is not a reason; "the corpus p10–p90 for display leading is 0.95–1.10 and we shipped 1.30" is. - Measure our own output the same way.
pnpm research:critiqueruns the identical in-page probe against pages the engine generated. Same instrument, same viewport, same scroll pass. - One widest gap per loop. Fix the lowest-scoring craft dimensions first. Breadth-first polishing hides regressions.
- Never let the score go down. If a loop lowers the craft score, revert the change or fix it before committing.
- Craft score is a floor, not a ceiling. A page can sit inside every band and still be dull. Each loop must also carry one qualitative judgement recorded in the ledger.
The loop
| Stage | Command / action | Output |
|---|---|---|
| 1. Goal | Write the loop goal into research/LOOP_LEDGER.md |
one sentence + target dimensions |
| 2. Widen | Add references to the local corpus if the current one is thin for the goal | research/corpus.local.json |
| 3. Measure | pnpm research:forensics |
research/measurements/ref-*.json |
| 4. Aggregate | pnpm research:aggregate |
research/aggregate.json, docs/10_DESIGN_EVIDENCE.md |
| 5. Critique | pnpm research:critique |
research/critique.json + weakest-dimension ranking |
| 6. Close the gap | Change the engine, tokens, or section grammar | code |
| 7. Re-critique | pnpm research:critique again |
new score, must be ≥ previous |
| 8. Record | Append the loop entry to research/LOOP_LEDGER.md |
ledger row |
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 · 149 lines · 70 tokens per session scan A 0fd78b95284e
design-research-loop is a skill published in the GitHub repository ashishpatill/tell-proof (2 stars, last pushed 22d ago), licensed MIT. It adds 70 tokens to every session and 1,777 once invoked, about $0.0003 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 2026-08-31.
Other skills, from other repositories
e2e-testing-patterns
Master end-to-end testing with Playwright and Cypress to build reliable test suites that catch bugs, improve confidence, and enable fast deployment. Use when implementing E2E tests, debugging flaky tests, or establishing testing standards.
playwright-recording
Record browser interactions as video using Playwright. Use for capturing demo videos, app walkthroughs, and UI flows for Remotion videos. Triggers include recording a demo, capturing browser video, screen recording a website, or creating walkthrough footage.
dev-browser
Browser automation with persistent page state. Use when users ask to navigate websites, fill forms, take screenshots, extract web data, test web apps, or automate browser workflows. Trigger phrases include "go to [url]", "click on", "fill out the form", "take a screenshot", "scrape", "automate", "test the website"…
browser-trace
Capture a full DevTools-protocol trace of any browser automation — CDP firehose, screenshots, and DOM dumps — then bisect the stream into per-page searchable buckets. Use when the user wants to debug a failed run, audit network/console/DOM activity, attach a trace to an in-progress session, or feed structured per-page…
tavily-usage
This skill should be used when user asks to "search the web", "fetch content from URL", "extract page content", "use Tavily search", "scrape this website", "get information from this link", or "web search for X".
Dead Link Detector
Automated detection and reporting of broken links, missing resources, and orphaned pages across web applications using crawl-based and DOM-based strategies.