Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add instructions/idavidov13/agentic-playwright/selectorsgit clone --depth 1 https://github.com/idavidov13/agentic-playwrightWhat 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 | $0.02174 | $0.02174 |
| Opus 5 | $0.01087 | $0.01087 |
| Sonnet 5 | $0.00435 | $0.00435 |
| Haiku 4.5 | $0.00217 | $0.00217 |
Grade A, and why
agentic-playwright selectors.instructions.md 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 2d 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 — 161 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Selector Strategy
Critical
- Selector priority order is mandatory:
getByRole>getByLabel>getByPlaceholder>getByText>getByTestId. Move to the next option only when the previous one is not feasible. - NEVER use XPath (
page.locator('//...')or'xpath=...'). - NEVER use CSS class or ID selectors as the primary strategy (
page.locator('.btn-primary'),page.locator('#submit')). Acceptable only as an absolute last resort after ruling out every semantic option. - Exploration with
playwright-cliis mandatory before writing any selectors. No guessing from wireframes, docs, or screenshots. Read theplaywright-cliskill for the commands. - If the app cannot be reached or auth fails, stop and notify the human — never ship placeholder locators with guessed names.
- String values inside
getByText(...)come fromenums/{area}/*(e.g.Messages.LOGIN_ERROR). Never hardcode repeated UI strings. See theenumsskill. - Every page object covering forms or CRUD must include feedback / validation message selectors — success, error, field validation, toast, loading, empty state as applicable. A page object without them is incomplete.
- Locators are
getaccessors returningLocator— this is a style/readability convention in the scaffold. Playwright'sLocatoris lazy, sogetand areadonlyfield set in the constructor behave identically at runtime.
Instructions
Phase 1: Open and authenticate
Never generate selectors from assumptions or documentation alone. Before writing any locators or page objects, explore the live application by running playwright-cli in the terminal (read the playwright-cli skill). Do not use IDE browser MCP, Cursor browser tools, or any substitute — orchestrator rule: No Substitute UI Exploration. If playwright-cli cannot run, stop and notify the human.
playwright-cli open <APP_URL>
playwright-cli snapshot
If the page fails to load or requires authentication:
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.
- 2d ago First seen · 161 lines · 2,174 tokens per session scan A 195cd7a4b298
agentic-playwright selectors.instructions.md is an instructions file published in the GitHub repository idavidov13/agentic-playwright (139 stars, last pushed 5d ago), licensed MIT. It adds 2,174 tokens to every session, about $0.0109 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 instructions, from other repositories
inspecto CLAUDE.md
Instructions for inspecto-dev/inspecto, covering inspecto — claude code development guide, project overview, monorepo structure, development phases (load each file as needed) and key architectural decisions.
qaskills CLAUDE.md
Instructions for PramodDutta/qaskills, covering claude.md, what this is, conventions (non-negotiable), commands and single package.
tapp AGENTS.md
AGENTS.md instructions for aarwitz/tapp, covering tapp agent playbook, no mcp connected? just run the cli, pick the right tool for the job, session driving (the playwright loop) and autonomous exploration (tappexplore).
vscode-web-components-ai AGENTS.md
Instructions for d13/vscode-web-components-ai, covering agents.md, project overview, common commands, prerequisites and development.
fodda-mcp CLAUDE.md
Claude Code instructions for piers-fawkes/fodda-mcp, covering fodda — claude code project instructions, two brains, two stores (do not confuse them), when piers asks you to prepare a brief — default procedure and repo topology & routing (agents hand off cross-repo).
vibe-testing CLAUDE.md
Claude Code instructions for AishwaryShrivastav/vibe-testing, covering vibe test — code-aware browser testing agent, what this project does, architecture, mcp tools (13) and development.