Skill: Research
Gather information about the user to inform portfolio content and design.
Sources (in priority order)
1. Materials Folder
Check /materials first:
materials/documents/ - Resume, cover letter, bio text
materials/images/ - Headshot, project screenshots
2. Profile Config
Read profile.yaml for:
- Name, contact info, social links
- Design preferences (sliders, archetype, inspirations)
- Content preferences (tone, focus, length)
- Personal notes (most valuable - user's own words)
3. GitHub
If username provided or discoverable:
- Pinned repositories
- README files for major projects
- Languages and contribution patterns
- Stars, forks (real metrics)
4. Web Search
Search for:
- "[name] software engineer [location]"
- "[name] github"
- "[name] projects"
- Company mentions, blog posts, talks
5. LinkedIn
Warning: Often blocked (error 999). If blocked:
- Skip entirely, don't retry
- Use web search for "[name] linkedin [company]" instead
- Check company About pages, press releases
Research Depth
Based on ai.research_depth in profile.yaml:
| Level |
Approach |
| 1-3 |
GitHub only, 1-2 search queries |
| 4-6 |
GitHub + web search + personal site (5-10 queries) |
| 7-8 |
Deep search, find hidden gems (15-20 queries) |
| 9-10 |
Exhaustive, find everything published (30+ queries) |
What to Extract
Factual (verify these)
- Current/past roles and companies
- Project names and tech stacks
- Education
- Public metrics (GitHub stars, npm downloads)
Contextual (ask if unclear)
- Impact and outcomes of work
- Specific responsibilities
- Scale (users, transactions, team size)
- Why they made certain choices
Asking Clarifying Questions
Use the AskUserQuestion tool for:
Experience gaps:
- "What did you work on at [Company]?"
- "What was the impact of [Project]?"
Project details:
- "What problem were you solving with [Project]?"
- "How many people use it?"
Direction:
- "What kind of role are you targeting?"
- "What do you want to emphasize?"
Question Guidelines
- Ask 3-5 questions at a time in logical groups
- Be specific, not generic ("What did you do at Uber?" not "Tell me about yourself")
- Don't ask about things clearly documented (dates, public GitHub stats)
Output
After research, you should know:
- Who this person is and what they do
- 3-5 key projects or achievements to highlight
- Their career trajectory and goals
- Their voice/tone from notes or writing samples
- What makes them unique (the hook)
If you can't answer these, ask more questions before proceeding.
1---2name: research3description: Gather information about users from materials, profile config, GitHub, and web search to inform portfolio content and design decisions.4---56# Skill: Research78Gather information about the user to inform portfolio content and design.910---1112## Sources (in priority order)1314### 1. Materials Folder15Check `/materials` first:16- `materials/documents/` - Resume, cover letter, bio text17- `materials/images/` - Headshot, project screenshots1819### 2. Profile Config20Read `profile.yaml` for:21- Name, contact info, social links22- Design preferences (sliders, archetype, inspirations)23- Content preferences (tone, focus, length)24- Personal notes (most valuable - user's own words)2526### 3. GitHub27If username provided or discoverable:28- Pinned repositories29- README files for major projects30- Languages and contribution patterns31- Stars, forks (real metrics)3233### 4. Web Search34Search for:35- "[name] software engineer [location]"36- "[name] github"37- "[name] projects"38- Company mentions, blog posts, talks3940### 5. LinkedIn41**Warning**: Often blocked (error 999). If blocked:42- Skip entirely, don't retry43- Use web search for "[name] linkedin [company]" instead44- Check company About pages, press releases4546---4748## Research Depth4950Based on `ai.research_depth` in profile.yaml:5152| Level | Approach |53|-------|----------|54| 1-3 | GitHub only, 1-2 search queries |55| 4-6 | GitHub + web search + personal site (5-10 queries) |56| 7-8 | Deep search, find hidden gems (15-20 queries) |57| 9-10 | Exhaustive, find everything published (30+ queries) |5859---6061## What to Extract6263### Factual (verify these)64- Current/past roles and companies65- Project names and tech stacks66- Education67- Public metrics (GitHub stars, npm downloads)6869### Contextual (ask if unclear)70- Impact and outcomes of work71- Specific responsibilities72- Scale (users, transactions, team size)73- Why they made certain choices7475---7677## Asking Clarifying Questions7879Use the `AskUserQuestion` tool for:8081**Experience gaps**:82- "What did you work on at [Company]?"83- "What was the impact of [Project]?"8485**Project details**:86- "What problem were you solving with [Project]?"87- "How many people use it?"8889**Direction**:90- "What kind of role are you targeting?"91- "What do you want to emphasize?"9293### Question Guidelines94- Ask 3-5 questions at a time in logical groups95- Be specific, not generic ("What did you do at Uber?" not "Tell me about yourself")96- Don't ask about things clearly documented (dates, public GitHub stats)9798---99100## Output101102After research, you should know:103- Who this person is and what they do104- 3-5 key projects or achievements to highlight105- Their career trajectory and goals106- Their voice/tone from notes or writing samples107- What makes them unique (the hook)108109**If you can't answer these, ask more questions before proceeding.**