digest

break down an external resource and identify what it means for us.

process

1. fetch the resource: $ARGUMENTS

   • use pdsx MCP for AT-URIs (bsky posts, leaflet docs, etc.)
   • use WebFetch for URLs
   • ask for clarification if the resource type is unclear

2. extract concrete information:

   • what are they doing?
   • what's their architecture/approach?
   • what are their constraints and priorities?
   • what's their roadmap?

3. cross-reference with our state:

   • check open issues for overlap or gaps
   • grep codebase for related implementations
   • identify where we align or diverge

4. identify actionable implications - the core output:

   • "given that X is true, we should consider Y"
   • specific issues to open or update
   • code changes to consider
   • integration opportunities
   • things we're missing or doing wrong

5. present findings - be direct:

   • lead with the implications, not the summary
   • include specific file:line or issue references
   • propose concrete next steps

anti-patterns

• philosophical musing without action items
• "we're complementary" without specifics
• agreeing that everything is fine
• backpedaling when challenged

output format

## implications

1. **[actionable item]**: [reasoning]
   - related: `file.py:123` or issue #456
   - suggested: [specific change or issue to create]

2. **[actionable item]**: ...

## extracted facts

- [concrete thing from the resource]
- [another concrete thing]

## open questions

- [things to clarify or investigate further]