Requirements
- Target platform
- OpenClaw
- Install method
- Manual import
- Extraction
- Extract archive
- Prerequisites
- OpenClaw
- Primary doc
- SKILL.md
Tencent SkillHub · Productivity
Secure Shopper
Asynchronous shopping research + checkout using secure-autofill (1Password-backed browser filling) with results recorded to workspace artifacts.
skill openclawclawhub Free
Asynchronous shopping research + checkout using secure-autofill (1Password-backed browser filling) with results recorded to workspace artifacts.
⬇ 0 downloads ★ 0 stars Unverified but indexed
Install for OpenClaw
Quick setup
- Download the package from Yavira.
- Extract the archive and review SKILL.md first.
- Import or place the package into your OpenClaw setup.
Package facts
- Download mode
- Yavira redirect
- Package format
- ZIP package
- Source platform
- Tencent SkillHub
- What's included
- SKILL.md, config.json, scripts/onboard.mjs, scripts/task_io.mjs
Validation
- Use the Yavira download entry.
- Review SKILL.md after the package is downloaded.
- Confirm the extracted package contains the expected setup assets.
Install with your agent
Agent handoff
Hand the extracted package to your coding agent with a concrete install brief instead of figuring it out manually.
- Download the package from Yavira.
- Extract it into a folder your agent can access.
- Paste one of the prompts below and point your agent at the extracted folder.
New install
I downloaded a skill package from Yavira. Read SKILL.md from the extracted folder and install it by following the included instructions. Tell me what you changed and call out any manual steps you could not complete.
Upgrade existing
I downloaded an updated skill package from Yavira. Read SKILL.md from the extracted folder, compare it with my current installation, and upgrade it while preserving any custom configuration unless the package docs explicitly say otherwise. Summarize what changed and any follow-up checks I should run.
Trust & source
Release facts
- Source
- Tencent SkillHub
- Verification
- Indexed source record
- Version
- 0.1.0
Provenance
- Publisher
- moodykong
- Source page
- View original listing
- Canonical URL
- Open canonical page
Documentation
secure-shopper 🛒
Find items across one or more shopping sites, summarize candidates, and (optionally) place the order using secure-autofill. This skill is asynchronous: spawn a sub-agent for browsing so the main chat stays responsive.
Required skills / plugin
The secure-autofill skill exists at: ~/.openclaw/skills/secure-autofill/ The secure-autofill plugin tools are available: vault_suggest vault_fill
Inherit secure-autofill prerequisites
A working non-headless Chrome (many shops block headless) Gateway environment has required env vars (per secure-autofill) Concrete check: command -v google-chrome || command -v google-chrome-stable
Configuration (portable)
Skill-local config files: Example (shareable, do not edit): ~/.openclaw/skills/secure-shopper/config.json.example Real (machine-specific, written by onboarding): ~/.openclaw/skills/secure-shopper/config.json Config keys: goToSites[]: list of default shopping sites (e.g. Amazon, Walmart) location.zip or location.address: used for shipping/availability context preferences.priority: one of: relevancy cheaper faster reviews preferences.maxCandidatesPerSite: cap per site (default 5) preferences.safeBrowsing: guardrails to avoid oversized pages / context overflow (applies to all sites) startFromSearch: true|false (default true) — prefer a site’s search results page over the homepage/product pages maxCandidatesPerPass: number (default 3) — extract a few items at a time (then paginate/scroll) snapshot: limits for browser.snapshot compact: boolean (default true) depth: number (default 6) maxChars: number (default 12000) fallback: what to do on context_length_exceeded retryWithTighterSnapshot: boolean (default true) switchToSearchUrl: boolean (default true)
Preferred (chat-first)
Ask Boss and then write config.json: Go-to shopping website(s) Examples: Amazon, Walmart, Target Store into goToSites[] Zip code OR proximity address Store into location.zip and/or location.address Preferences Ask for priority: relevancy vs cheaper vs faster delivery vs higher review scores Store into preferences.priority (+ optional notes) After collecting answers, update the real config file. Optional helper (terminal): node ~/.openclaw/skills/secure-shopper/scripts/onboard.mjs \ --sites 'Amazon=https://www.amazon.com|Walmart=https://www.walmart.com' \ --zip 46202 \ --priority cheaper
0) Require a shopping description
The user must provide a description of their shopping task. If they didn’t: stop and ask for it.
1) Honor runtime user prompts
Runtime user instructions (the user’s message for this run) override stored config. Examples of runtime overrides: “Use Target instead of Amazon.” “Only show Prime-eligible.” “Budget under $50.”
2) Login via secure-autofill (skip if already logged in)
Use the configured go-to sites, unless the runtime prompt specifies a site. If the site session appears already authenticated: skip login. Otherwise, use secure-autofill login flow: browser.snapshot to get refs vault_suggest/vault_fill to fill credentials
3) Make the browsing asynchronous
Immediately after accepting the task, respond with something like: I’m en route to the stores. I’ll notify you when I find the best matches. Then spawn a sub-agent so the main session is not interrupted. Implementation note: Use sessions_spawn with a task that includes the shopping description and any runtime overrides.
4) Browse + identify candidates
The sub-agent browses each chosen site, searches, filters, and identifies candidates that fit the user description. Context-safe browsing (ALL shopping sites) Many shopping sites can produce extremely large pages/snapshots. To avoid context_length_exceeded failures: Prefer starting from a search results URL (or the site’s search box) rather than the homepage. Use small snapshots: browser.snapshot(..., compact=true) keep depth modest (e.g., 4–8) set maxChars and/or target a specific container when possible Extract incrementally: grab top ~3 candidates, record them, then paginate/scroll and repeat until maxCandidatesPerSite is met If a snapshot still overflows: retry with a tighter snapshot (smaller depth / smaller region) switch to a search URL (/search?q=...) and re-extract Do not “reason through” massive dumps. If the page is huge, reduce the page slice first. Record results to: /home/miles/.openclaw/workspace/artifacts/secure_shopping/{timestamp}_shopping_task.json JSON requirements: Record: userPrompt (shopping description) startTime endTime phase (required): candidates_found | awaiting_accept_deny | awaiting_checkout_confirm | ordered candidates[] Candidates for the same request must live under the same parent task. Each candidate must include: price (string) reviewScore (string/number) url verdict (short) status: pending | accepted | denied | shopped Suggested candidate shape: { "site": "Amazon", "title": "...", "price": "$39.99", "reviewScore": "4.6 (12,345)", "url": "https://...", "verdict": "Best value under $50; good reviews; ships tomorrow", "status": "pending" } Helper module (optional): scripts/task_io.mjs.
5) Notify user + REQUIRE accept/deny (hard gate)
When browsing is done, you must: Set JSON phase = "awaiting_accept_deny". Translate the JSON into a human-friendly summary. In the same message, require an ACCEPT/DENY decision. Do not end the turn without the prompt. Mandatory message template (copy this structure): Recommended pick: <title> — <price> — <reviewScore> — <1-line why> Other options: (optional, 1–5 bullets) Choose: Reply with A=accept/deny, B=accept/deny, ... (or “Accept A” / “Deny B”). Next step: “If you accept one: do you want me to checkout, or stop at ready-to-buy?” Hard rule: If you listed candidates/links but did not include an explicit Choose (ACCEPT/DENY) line, the output is invalid and must be rewritten before sending.
6) Apply accept/deny updates
Once the user replies: Update each candidate status to accepted or denied. Confirm the accepted candidate(s). Set JSON phase: awaiting_checkout_confirm if at least one is accepted and checkout is not yet confirmed keep awaiting_accept_deny if the user’s response is ambiguous / incomplete
7) Checkout (only after explicit confirmation)
Before you click any “Place order” / “Submit” equivalent: Ask for a clear confirmation like: “Confirm checkout for A? (yes/no)” Set JSON phase = "awaiting_checkout_confirm" until confirmed. If the user confirms checkout: Navigate to the accepted candidate’s URL Add to cart / proceed to checkout Use secure-autofill to input payment/shipping info and submit If secure-autofill reports an error: Do not guess. Pass the error back to the user.
8) Mark as shopped
If the order is successfully placed: update that candidate’s status to shopped set JSON phase = "ordered"
Notes / guardrails
Never paste secrets. Checkout flows often require MFA / SMS verification; ask the user when needed. Prefer fewer high-quality candidates over a long list.
Category context
Workflow acceleration for inboxes, docs, calendars, planning, and execution loops.
Source: Tencent SkillHub
Largest current source with strong distribution and engagement signals.
Package contents
Included in package
2 Scripts1 Docs1 Config
- SKILL.md
- scripts/onboard.mjs
- scripts/task_io.mjs
- config.json