agenticnotetaking/skills/research-recursive

/eidos:research-recursive

Recursive research on a broad topic where each area gets its own file. Areas can branch into sub-areas as nested wiki-linked research files, creating a navigable tree.

Usage

/eidos:research-recursive [topic or domain]                  # start — will ask for budget
/eidos:research-recursive [topic] budget 10                  # start with soft budget of 10
/eidos:research-recursive [topic] budget 10 hard             # start with hard budget of 10
/eidos:research-recursive [topic] budget none                # no budget constraint
/eidos:research-recursive continue                           # resume existing recursive research

When to Use

Use research-recursive (not research-deep) when:

• The topic is broad enough that areas deserve their own files
• You want to branch arbitrarily deep on promising threads
• The research tree should be navigable — readers pick which branches to follow
• Each area may grow large enough that one file would become unwieldy

Use /eidos:research-deep when all areas belong in a single document. Use /eidos:research for focused questions with a specific answer.

Instructions

Mode Detection

Check ARGUMENTS:

• If continue → go to Continue Mode
• Otherwise → go to Start Mode (ARGUMENTS is the topic)
• Parse budget N and optional hard from args if present

---

Start Mode

1. Clarify and Scope

Use AskUserQuestion to establish:

• What's the overarching question or domain?
• Any known areas of interest, or should we discover them?
• How deep should we go? (survey vs. deep dive)
• Any specific angles, constraints, or contexts?

Budget: If not provided in args, ask:

• How many areas (budget)? Each area file — top-level or nested — costs 1 unit.
• Hard or soft? Soft = agent can exceed with justification or ask for more. Hard = strict cap.
• "None" is valid — no budget constraint, agent uses judgement on depth.

2. Create Root File

Think about the topic and identify 4-8 areas that together cover the landscape.

Run date '+%y%m%d%H%M' to get the current timestamp. Create memory/research-recursive - <timestamp> - <claim>.md — this is the root file.

---
tldr: Brief description of the research topic
status: active
budget: 10
budget_type: soft
budget_spent: 0
---

# Research: [Topic]

## Meta

This is a recursive research tree.
Each area is explored in its own file, linked from this root.
Areas can branch deeper — sub-areas become nested files linked from their parent.
Budget is shared across the entire tree.

## Seed

> [Verbatim user intent that triggered this research — preserved as-is]

## Question

The overarching question or domain to explore.

## Areas

1. [ ] [[research-recursive - <timestamp> - <area 1 claim>]] — brief description
2. [ ] [[research-recursive - <timestamp> - <area 2 claim>]] — brief description
3. [ ] [[research-recursive - <timestamp> - <area 3 claim>]] — brief description
...

## Trajectory Adjustments

<!-- When earlier areas change the research direction, document it here with timestamps. -->

## Synthesis

<!-- Cross-area conclusions — written after areas are explored, or updated incrementally. -->

## Implications

<!-- How this affects current work. Specs to create, decisions to make, concepts to adopt. -->

Do not create the area files yet — just list them as wiki links in the root. Commit the root file.

Present the area outline:

Research plan for [topic]:

1. [Area 1] — brief description
2. [Area 2] — brief description
...

Budget: N [soft|hard|none]
Starting with Area 1.
continue?

3. Research Areas

For each area:

1. Run date '+%y%m%d%H%M' for the area file timestamp
2. Create memory/research-recursive - <timestamp> - <area claim>.md:

---
tldr: Brief description of this area
status: active
parent: "[[research-recursive - <root timestamp> - <root claim>]]"
---

# [Area Name]

## Research Plan

1. [ ] First thing to investigate
   - context or approach notes
2. [ ] Second thing to investigate
3. [ ] ...

## Sources

### [Source name / URL]

- Key points
- Relevant detail
- Caveats

## Findings

Local synthesis for this area — what did we learn?

## Threads

- Sub-questions or areas that emerged
- => spawned [[research-recursive - <timestamp> - <sub-area claim>]]

3. Work through the Research Plan — investigate, add sources, write findings
4. Note threads — sub-questions that emerged
5. If a thread warrants its own area and budget allows: create a nested area file, link it from Threads with =>, and add it to the parent's Areas list
6. Set status to done
7. Mark [x] in the parent file's area list
8. Increment budget_spent in the root file's frontmatter
9. Commit

Budget awareness — after completing each area, check budget in the root file:

• Under budget: proceed normally. Note threads that could branch but weigh against remaining budget.
• Approaching budget (>75% spent): be selective. Only branch for genuinely important threads. Note skipped threads with => would explore with more budget: [description].
• At budget (soft): ask the user — "I've used N/N budget. [Thread X] seems important. Can I get M more units?"
• At budget (hard): stop exploring new areas. Finish current area, then move to Synthesis.
• No budget: use judgement.

After each area, give a brief status update:

Area done: [one-line summary]
Budget: N/M spent [soft|hard]
Threads: [any new branches created]
Moving to next area.

After every 2-3 areas, pause:

Progress: X/Y areas complete. Budget: N/M spent.
[Brief summary of what's emerging]
continue?

4. Adjust Trajectory

If earlier areas reveal that the research direction needs to change:

• Add/remove areas in the root file
• Update research plans in upcoming area files
• Document the change in the root's Trajectory Adjustments section with timestamp
• Ask the user before adding more than 2 new top-level areas

5. Synthesise

After all areas are explored:

1. Write the Synthesis section in the root file — patterns, tensions, gaps, frameworks
2. This is cross-area thinking, not area summaries
3. Write Implications — connect to current work
4. Set status to complete
5. Commit

6. Present and Offer Next Steps

Research complete: [[research-recursive - <timestamp> - <claim>]]

[Key synthesis points]

Tree: N files (root + M area files)

Options:
1 - Act on implications (create specs, decisions, etc.)
2 - Go deeper on [specific area or thread]
3 - Done for now

---

Continue Mode

1. Find Research

Search memory/ for research-recursive - *.md files that have no parent in frontmatter (these are root files) and status not complete or superseded.

Single active root: load it. Multiple active: present a selection list. None active: offer to start a new one.

2. Read and Assess State

Read the root file. For each area listed, check if the linked file exists and read its status. Determine:

• Which areas are done, which are open
• Whether the current area has open research plan items
• What threads were noted but not yet branched
• Whether trajectory adjustments are needed

3. Resume

Present state and propose next action, then continue with the relevant step from Start Mode.

---

Principles

One file per area. Each area is self-contained — a reader can open one area file and understand it without reading the rest. The root file is the map; area files are the territory.

Wiki links are the structure. The tree emerges from [[wiki links]] between files, not from heading hierarchy. This means areas can be reorganised by updating links, not restructuring a monolith.

Budget is global. All area files — regardless of nesting depth — draw from the same budget in the root file. This forces the same breadth-vs-depth trade-off as research-deep.

Depth follows value. Don't branch just because you can. A thread becomes a sub-area file only when it has enough substance to warrant its own research cycle.

Output

• Creates: memory/research-recursive - <timestamp> - <claim>.md (root)
• Creates: memory/research-recursive - <timestamp> - <area claim>.md (per area)
• Nested areas create further files linked from their parent
• Root status starts as active, set to complete when done
• Intermediate commits after each area file