Reading List Prioritizer

Reading List Prioritizer

This skill takes your personal reading list (a local markdown file with article titles and/or URLs) and returns a structured weekly reading plan: items scored by relevance to your role, grouped by topic, with a "Read this week" shortlist at the top.

Input:

• A markdown file (.md) with your reading list — one item per line, as [Title](URL) links, bare URLs, or plain text titles
• Your role or current focus area (stated in 1–2 sentences in your message)

Output:

• Markdown response with a prioritized "Read this week" shortlist + full list grouped by topic, each item labeled [High] / [Med] / [Low]

Language Detection

Detect the user's language from their message:

• If Russian (or contains Cyrillic): respond in Russian
• If English (or other Latin-script language): respond in English
• If ambiguous: respond in the language of the trigger phrase used

Instructions

Step 1: Validate Input

1. Check that a reading list file has been provided (uploaded or pasted)

   • If no file and no pasted list: stop and return: "Please provide your reading list as a .md file or paste the list directly."
   • Accepted formats: .md file, pasted markdown list, plain text list

2. Check that the user's role or focus area is stated in their message

   • If missing: ask for it before proceeding — "What's your current role or focus area? (e.g., 'PM focused on AI product strategy')"
   • Do not guess or assume a generic role — prioritization depends on it

3. Validate file content: must contain at least 3 parseable items

   • If fewer than 3 items: proceed but note "Short list — showing all items with priority labels"
   • If file is empty or contains no recognizable items: stop with message "The reading list appears empty. Please check the file format."

Step 2: Parse the Reading List

1. Read each line of the list and extract:

   • Title: from [Title](URL) format, or plain text line, or domain-derived label if only a bare URL
   • URL/domain: extract domain from URL if present (e.g., nngroup.com, hbr.org, stratechery.com)
   • Tags or notes: if the user appended a tag or comment after the link (e.g., [Article](url) #ux), extract them

2. Build a working list of items: [title, domain, tags] per item

3. Silently remove exact duplicate URLs; note count at the bottom of output if any found

Step 3: Detect Topics

1. Analyze all item titles and domains together to identify 2–7 topic clusters

   • Use title keywords and known domain signals:
     • nngroup.com, smashingmagazine.com → UX/Design
     • hbr.org, mckinsey.com → Leadership/Strategy
     • stratechery.com, ben-evans.com → Business/Tech Strategy
     • towardsdatascience.com, arxiv.org → AI/Data
     • substack.com, newsletters → Industry News
   • For unknown domains: use title keywords to assign topic
   • Items that don't fit any cluster: assign to "General"

2. Aim for 2–7 distinct groups; if all items cluster into one topic, add subtopic labels where possible

Edge Cases:

• All items from a single domain: group by subtopic using title keywords; note "All items from one source — grouped by subtopic"
• Items in mixed languages (EN + RU titles): process both; group by topic regardless of language
• Items with only bare URLs and no titles (domain cannot be identified or title keywords unavailable): assign to "General"; add note at top of Full List section: "Note: grouping is approximate — limited metadata available for these items"

Step 4: Score Relevance

1. For each item, score relevance to the user's stated role/focus area:

   • [High]: Directly relevant — topic matches role or stated focus area, or title addresses a specific problem the user likely faces
   • [Med]: Broadly relevant — adjacent topic, useful background, or industry-relevant but not core to stated focus
   • [Low]: Low relevance — interesting but tangential to stated role and focus area

2. Apply scoring across all items before ranking

3. Within each topic group, sort: High → Med → Low

Step 5: Build "Read This Week" Shortlist

1. Select top 5–7 items overall:

   • Prefer [High] items from the widest topic variety (1–2 per topic max in shortlist)
   • If fewer than 5 [High] items exist, include top [Med] items

2. For each shortlisted item, add a 1-sentence relevance note explaining why it's recommended

3. Cap at 7 items regardless of list size

Edge Cases:

• Very long list (50+ items): process all; limit shortlist to 7; add note "Remaining X items are in the full grouped list"

Step 6: Format and Output

1. Build the output markdown (see Output Format below)
2. Include all items from input in the full grouped list — no items silently dropped
3. Add footer with count of duplicates removed (if any)

Negative Cases

• Empty file: Return "The reading list file appears empty or has no recognizable items. Paste your list directly or check the file format."
• Non-md / unreadable file: Return "Expected a markdown (.md) file with a list of articles or links. Please convert your list to markdown format."
• No role/focus provided: Ask once before proceeding — do not produce output without it (prioritization would be meaningless without context)
• All items unresolvable (no titles, no domains, no text): Return "Could not parse any items from the list. Check that each line contains a title or URL."

Output Format

## Reading Plan — [YYYY-MM-DD]
**Focus area:** [user-stated role and focus]
**Total items:** X | **Duplicates removed:** Y (or omit if 0)

---

### Read This Week (Top 5–7)
1. [Title](URL) — [Topic] · [1-sentence relevance note]
2. [Title](URL) — [Topic] · [1-sentence relevance note]
...

---

### Full List by Topic

#### [Topic 1 name]
- [High] [Title](URL)
- [Med] [Title](URL)
- [Low] [Title](URL)

#### [Topic 2 name]
- [High] [Title](URL)
...

---

### Skippable This Week
- [Low] [Title](URL) — low relevance to stated focus
...

Field rules:

• "Read This Week" items: include relevance note (why recommended for this role)
• Priority labels: always present, always one of [High] / [Med] / [Low]
• "Skippable This Week": list only [Low] items; include a brief reason if non-obvious
• If all items are [High] relevance: omit "Skippable This Week" section
• Bare URLs without titles: display as domain name (e.g., stratechery.com — [article path])