id: competitive-intel-digest title: Competitive-Intel Digest category: research primitive: routine use_when: You want a scheduled digest of changes to a watchlist of competitor pages, with each claim linked to its source. verify: Every digest item links to the specific source page and quotes the changed text; nothing is unsourced. tags: [monitoring, digest, routine]

Competitive-Intel Digest

A scheduled routine that diffs a competitor watchlist and digests only what changed, each item linked to its source. Primitive: routine · Category: research

Use when

You track a handful of competitor pages (pricing, changelog, docs) and want a periodic digest of real changes — not a re-summary of unchanged pages — with every claim traceable to the page it came from.

Prompt

Goal:        Produce a <cadence> digest of CHANGES to the watchlist pages.
Context:     The watchlist URLs; the previously captured snapshot of each.
Constraints: Report only changed content. Every item must link + quote the source.
Done-when:   Every changed page has a digest item; unchanged pages are omitted.
Evidence:    A digest (item, source URL, quoted change); the snapshot diff per page.
If-blocked:  If a page is unreachable, note it as "could not fetch", don't infer changes.

Verify

A separate check confirms each digest item maps to a page whose snapshot actually changed, and that the quoted text appears on the source page.

Steps

1. Re-fetch each watchlist page; diff against the prior snapshot.
2. Digest only the changed pages, quoting + linking the change.
3. Save new snapshots for next run.

Notes

This is a routine (durable, scheduled), not a session /loop — it should keep running after you log off. Weekly or daily is plenty; sub-hourly just re-fetches unchanged pages.