competitive-pricing
Research competitor pricing strategies by extracting pricing pages, comparing product listings on Google Shopping and Amazon, and analyzing market positioning. Produces pricing benchmarks, tier comparisons, and feature-per-price analysis.
No fetch-and-execute, destructive, credential, exfiltration or rules-override patterns in 2 files, checked on Oct 20, 2026. This is not a guarantee of safety. How the scan works.
npx degit sandbaseai/sandbase-skills/research/competitive-pricing .claude/skills/competitive-pricing12 alternatives · Best data & science skills · Author? Prove this skill
SKILL.md
# Competitive Pricing Research competitor pricing strategies by extracting pricing pages, comparing product listings on Google Shopping and Amazon, and analyzing market positioning. Produces pricing benchmarks, tier comparisons, and feature-per-price analysis. Read [the API map](references/sandbase-api-map.md) before selecting a capability. ## Call SandBase capabilities Use the capability identifiers below as discovery hints, not MCP tool names. Find the matching endpoint with `sandbase_discover(q: "<provider and capability>")`; use its returned `name` in `sandbase_inspect(name: "<returned name>")`. Read `inputSchema`, pricing, and `execute_as`, then call `sandbase_run` using `execute_as.arguments.name` and schema-defined `arguments`. If a `run_id` is returned, poll `sandbase_run_get(run_id: "<returned run_id>")` within the task budget until `completed` or `failed`; report pending or failed runs without resubmitting them automatically. ## Capability identifiers - `strale_pricing_page_extract` - `google_shopping_bulk_products` - `amazon_bulk_product_details` - `context_dev_extract_structured_data` - `firecrawl_scrape` ## Workflow 1. Understand the user's research question, target, and context. 2. Resolve each selected capability with `sandbase_discover`, then inspect the returned `name` with `sandbase_inspect` to confirm the schema and pricing. 3. Follow `execute_as` with `sandbase_run`; collect async results with `sandbase_run_get` as described above. 4. Synthesize findings into a clear, evidence-backed answer. 5. Cite sources, note evidence gaps, and separate observations from interpretations. ## Guidelines - Always call `sandbase_inspect` before using any capability. - Cite sources and preserve attribution (URLs, usernames, dates, metrics). - Separate factual observations from analysis and recommendations. - If data is unavailable, note the gap and continue with available evidence. - Read-only research only. Never take actions on platforms.
Shown under the source repo’s Apache-2.0 license. Read the full folder on GitHub.