competitor-page-watcher
Watches specific web pages and reports the day anything changes — pricing, job postings, product updates. The user adds a page with "watch: [URL] — [label]"; the agent snapshots it to memory and re-checks daily, alerting on Telegram with exactly what moved (old price → new price, added/removed plans, new job postings).
ソース情報
- リポジトリ
- nearai/ironhub
- ソースの最終更新活動
- 2026年6月22日 13:54
- 検出された SKILL.md の言語
- 英語
- スター
- 8
- フォーク
- 14
インストール方法
デフォルトでは、最初にソースを確認する Prompt が選択されています。直接コマンドに切り替えるか、ローカルコピーをダウンロードすることもできます。
ソースファイルを確認
インストールを決める前に、SKILL.md と SkillsMP に表示されている付属ファイルをお読みください。
SKILL.md を表示中
SKILL.md
ソースの指示 · 読み取り専用プレビュー- name
- competitor-page-watcher
- version
- 1.0.0
- description
- Watches specific web pages and reports the day anything changes — pricing, job postings, product updates. The user adds a page with "watch: [URL] — [label]"; the agent snapshots it to memory and re-checks daily, alerting on Telegram with exactly what moved (old price → new price, added/removed plans, new job postings).
- activation
- {"keywords":["watch:","competitor","watched pages","stop watching","pricing page","track competitor"],"patterns":["(?i)watch:\\s*https?://","(?i)(show|list)\\s+watched\\s+pages","(?i)stop\\s+watching","(?i)check\\s+https?://.*\\s+now"],"tags":["monitoring","competitive-intelligence","automation"],"max_context_tokens":2000}
- requires
- {"tools":["memory","time","http","message","routine"],"bins":[],"env":[]}
You watch web pages and tell the user the day anything meaningful changes — pricing, job postings, product updates.
## Hard rules
- Only watch pages the user explicitly added. Never start watching a URL on your own.
- Only fetch publicly accessible pages. Never log in, bypass a paywall, or submit a form to reach content, and never store credentials or session data in memory.
- Report only changes confirmed by comparing the fresh fetch against the saved snapshot. Never invent, exaggerate, or guess a change.
- If a fetch fails (page down, blocked, timeout), report the failure. Never treat a failed fetch as "no change."
- The "what this might signal" line is interpretation, not fact — one line, framed as a guess.
- Send a Telegram alert only when something actually changed. Otherwise reply `HEARTBEAT_OK` and stop; never send a "nothing changed" message.
## Adding a page
When the user says `watch: [URL] — [what it is]` (e.g. `watch: https://competitor.com/pricing — their pricing page`):
1. Fetch the URL with the `http` tool.
2. Pull out the meaningful text — ignore navigation, footers, scripts.
3. Read `competitors/pages.md` with `memory_read`.
4. Save: URL, label, a summary of the current content, key facts (prices, plan names, job titles), date captured. Write it back with `memory_write`.
5. Confirm: "Now watching [label]. I'll check daily and alert you on any change."
## Daily check (routine)
Create a routine that runs every day at 8:00 AM:
1. Read `competitors/pages.md`.
2. For each watched page: fetch it, pull the meaningful text, compare against the saved version.
3. If it changed, identify exactly what moved — price changes (old → new), added/removed plan tiers, added/removed job postings, new product features.
4. Update the saved version in memory.
5. If anything changed, send the Telegram alert below. If nothing changed, reply `HEARTBEAT_OK` and stop.
Alert format:
🔍 Competitor Change Detected — [date]
📄 [Competitor pricing page]
- Pro plan: $49/mo → $59/mo (price increase)
- New tier added: 'Enterprise' at $199/mo
📄 [Competitor careers page]
- New posting: 'Head of Sales' — they're building a sales team
- Removed: 'Junior Developer'
💡 What this might signal: [one line per change]
## Commands
- "show watched pages" — list all monitored URLs with their last-change date
- "stop watching [URL]" — remove a page
- "check [URL] now" — force an immediate check of one page
GitHubで見る