agentsclimarketplace

Lenny research

Skill mnttnm/claude-skills/lenny-research

Custom Claude Code skills spanning UI/UX, productivity, and developer tooling

Install
npx -y skills add mnttnm/claude-skills --skill lenny-research

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

  • no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
  • 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Research Lenny Rachitsky's archive of 349 newsletter posts and 289 podcast interviews for comprehensive, well-cited answers. Auto-trigger when the user asks practical questions about startups, product management, growth, B2B SaaS, pricing, leadership, career development, hiring, workplace soft skills, AI product work, go-to-market, analytics, org design, or user research. Handles topic deep-dives, guest-specific queries, actionable advice requests, career transition questions, cross-topic comparisons, and multi-faceted questions spanning several domains. Use the lennys-data MCP tools for search and retrieval.

SKILL.md

3.5 KB, as published. Nobody here has run it

Lenny's Archive Research

Research Lenny's archive using multi-probe retrieval for comprehensive, well-cited answers.

Workflow

1. Decompose the Query

Identify query type and search dimensions:

  • Topic deep-dive: "What has Lenny covered about pricing?" → single topic, broad search
  • Guest-specific: "What did Shreyas Doshi say about PM craft?" → guest + topic
  • Actionable advice: "How should I run my first 1:1?" → practical frameworks
  • Comparison: "Compare advice on PMF from different guests" → multiple perspectives
  • Multi-faceted: "Moving from engineering to product in the AI era" → 2-3 topic dimensions

Read references/topic-map.md to expand query terms into pipe-delimited search probes. Most real questions span 2-3 topic clusters — create a separate search probe for each dimension.

2. Parallel Multi-Probe Search

Fire 2-3 search_content calls in parallel with different term expansions. Use content_type filter only when query clearly targets one type.

Example for "Moving from engineering to product in the AI era":

Probe 1: search_content("career|transition|role|engineering to product|switch", limit=15)
Probe 2: search_content("AI|LLM|product manager|new world|changing", limit=10)

If Probe 1 returns <3 results, skip further probes — topic is sparse, note this in response.

3. Merge, Rank, Extract

Deduplicate by filename. Rank: multi-probe overlap > title match > tag relevance > recency.

For top 5-10 results, extract in parallel:

  • Newsletters (typically <5K words): read_content for full text
  • Podcasts (typically >10K words): read_excerpt with radius=800, iterate match_index 0-2

See references/retrieval-guide.md for detailed extraction strategy per query type.

4. Synthesize

Adaptive depth based on query type:

Query typeSourcesStyle
Topic deep-dive5-10Group by sub-theme, multiple perspectives
Guest-specific3-5 episodesOrganize by topics discussed
Actionable advice3-5Lead with clearest framework, then alternatives
Comparison3-5 per sideTable or side-by-side format
Quick factual1-3Direct answer with citation

Citation Rules

  • Cite every claim: [Post Title](filename) — include guest name for podcasts
  • Include 2-3 direct quotes per response when pithy/memorable quotes exist
  • Group findings by sub-theme, not by source file
  • Flag sparse coverage: "I found limited coverage of this in Lenny's archive"
  • Note: the archive excludes the last 3 months of newsletter posts

Speed Rules

  • All search probes execute in parallel (single tool call batch)
  • All excerpt/content reads execute in parallel
  • Never full-read a podcast transcript unless specifically needed
  • Skip secondary probes when primary search returns <3 results

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.