Searches and retrieves past learnings from the Learning Log by topic, category, or date. Activates when the user wants to find past insights, search their learning history, or asks 'what did I learn about [topic]?' Supports keyword search, category filtering, date ranges, and source-based browsing.
From founder-osnpx claudepluginhub thecloudtips/founder-os --plugin founder-osThis skill uses the workspace's default tool permissions.
references/search-filter-logic.mdDesigns and optimizes AI agent action spaces, tool definitions, observation formats, error recovery, and context for higher task completion rates.
Enables AI agents to execute x402 payments with per-task budgets, spending controls, and non-custodial wallets via MCP tools. Use when agents pay for APIs, services, or other agents.
Compares coding agents like Claude Code and Aider on custom YAML-defined codebase tasks using git worktrees, measuring pass rate, cost, time, and consistency.
Query, filter, and browse past learnings from the Learnings Notion database. Used by: /founder-os:learn:search.
Provide fast, relevant access to previously captured learnings. Support filtering by topic, keyword, date range, and source type. Return ranked results with related-insight cross-references for each match. This skill operates in read-only mode against the discovered Learnings database (see Notion Database Access below).
Database name: Search for [FOS] Learnings first, then fall back to Founder OS HQ - Learnings, then Learning Log Tracker - Learnings.
Read-only access. Search the workspace for a database titled "[FOS] Learnings". If not found, try "Founder OS HQ - Learnings". If not found, fall back to "Learning Log Tracker - Learnings". If none exists, report that no learnings have been logged yet and suggest using /founder-os:learn:log to start.
| Property | Type | Used in Search |
|---|---|---|
| Title | title | Text search, display |
| Insight | rich_text | Text search, display |
| Topics | multi_select | Topic filter, display |
| Source Type | select | Source filter |
| Context | rich_text | Text search |
| Related IDs | rich_text | Cross-reference lookup |
| Related Titles | rich_text | Display |
| Week | rich_text | Date-range filter |
| Logged At | date | Date-range filter, sort, display |
Apply filters in this fixed order: topic → date → keyword. Each filter narrows the result set for the next.
If the positional argument or --topic flag specifies a topic name, filter to entries where Topics contains the specified topic (case-insensitive match). Accept partial matches: "tech" matches "Technical".
If the positional argument looks like a general search term rather than a topic name (not matching any of the 10 predefined topics), skip this filter and treat it as a keyword instead.
If --since=Nd or --since=YYYY-MM-DD is specified, filter to entries where Logged At is on or after the resolved date.
--since=7d → last 7 days from today--since=30d → last 30 days--since=2026-01-15 → on or after January 15, 2026--since flag: no date filter (return all time)If a keyword is provided (either as the positional argument when not matching a topic, or via additional text), filter to entries where Title, Insight, or Context contains the keyword string. Apply case-insensitive substring matching.
After filtering, rank results using a composite relevance score. Calculate the score for each result by summing these points:
| Signal | Points |
|---|---|
| Keyword match in Title | +3 |
| Keyword match in Insight | +2 |
| Keyword match in Context | +1 |
| Exact topic match (topic filter active) | +2 |
| Recency: logged in last 7 days | +1 |
| Recency: logged more than 90 days ago | -1 |
Sort results by composite score descending. Tiebreak by Logged At descending (most recent first).
Present results as a numbered list with topic emoji tags and key metadata:
### 🔍 Learning Search Results
**Query**: [search terms] | **Filters**: [active filters] | **Results**: N found
1. 📘 **[Title]** — Technical, Process
_[First 120 chars of Insight text]..._
📅 [Logged At date] | Source: [Source Type]
🔗 Related: [Related Titles or "none"]
2. 💡 **[Title]** — Idea, Business
_[First 120 chars of Insight text]..._
📅 [Logged At date] | Source: [Source Type]
🔗 Related: [Related Titles or "none"]
| Topic | Emoji |
|---|---|
| Technical | 📘 |
| Process | ⚙️ |
| Business | 💼 |
| People | 👥 |
| Tool | 🔧 |
| Strategy | 🎯 |
| Mistake | ⚠️ |
| Win | 🏆 |
| Idea | 💡 |
| Industry | 🌐 |
Use the emoji of the first topic in the entry's Topics list as the leading icon.
Default: 10 results. Override with --limit=N. Maximum: 50 results per query. If more results exist beyond the limit, note the total count and suggest narrowing the search.
For each search result, display the Related Titles value if non-empty. This provides cross-referencing without additional database queries. If Related Titles is empty, display "none".
When no results match the query:
--since is activeIf none of the [FOS] Learnings, Founder OS HQ - Learnings, or Learning Log Tracker - Learnings databases exists, report clearly: "No learning log found. Use /founder-os:learn:log to capture your first learning." Do not create the database on read operations.
Report the connection error and suggest retrying. Do not fabricate search results.
When the first argument could be either a topic name or a keyword (e.g., "process" matches the Process topic but could also be a keyword), prioritize topic matching. If the user seems to want keyword search instead, suggest using --topic=none or rephrasing.
For queries with no topic filter, no date filter, and a single common word as keyword, warn that results may be broad and suggest adding a topic or date filter.
For detailed filter construction and Notion query patterns, consult:
${CLAUDE_PLUGIN_ROOT}/skills/learn/learning-search/references/search-filter-logic.md — Sequential filter pipeline construction, Notion API query building, composite relevance scoring formula, and edge case handling for complex multi-filter queries