| name | exps - Exploration |
| description | Systematically explores an unfamiliar space by mapping boundaries, identifying dimensions, sampling broadly, and flagging surprises and promising regions. |
| output | {"format":"prose"} |
Exploration
Input: $ARGUMENTS
Step 1: Map the Boundaries
Define the edges of the space before exploring the interior.
SPACE DEFINITION:
- Domain: [what area are we exploring?]
- In scope: [what's included]
- Out of scope: [what's excluded and why]
- Known unknowns: [things we know we don't know]
- Adjacent spaces: [related areas we're not exploring yet]
Rule: If you can't define the boundaries, the space is too vague. Narrow it.
Step 2: Identify Dimensions
Find the axes along which this space varies.
DIMENSIONS:
1. [Dimension 1]: [what it measures] — range: [low end] to [high end]
2. [Dimension 2]: [what it measures] — range: [low end] to [high end]
3. [Dimension 3]: [what it measures] — range: [low end] to [high end]
4. [Dimension 4]: [what it measures] — range: [low end] to [high end]
Aim for 3-6 dimensions. Fewer means you're missing structure. More means some dimensions are redundant.
Step 3: Sample Broadly
Take a wide sweep before going deep anywhere.
BROAD SAMPLE:
| Region | Description | Initial Assessment | Promising? |
|--------|------------|-------------------|------------|
| [region 1] | [what's here] | [first impression] | [yes/maybe/no] |
| [region 2] | [what's here] | [first impression] | [yes/maybe/no] |
| [region 3] | [what's here] | [first impression] | [yes/maybe/no] |
| [region 4] | [what's here] | [first impression] | [yes/maybe/no] |
| [region 5] | [what's here] | [first impression] | [yes/maybe/no] |
| [region 6] | [what's here] | [first impression] | [yes/maybe/no] |
Rule: Sample at least 5 regions. Resist the pull to dive deep into the first interesting thing.
Step 4: Track Coverage
Monitor what's been explored and what hasn't.
COVERAGE MAP:
- Explored: [list of regions/dimensions covered]
- Unexplored: [list of regions/dimensions not yet touched]
- Coverage estimate: [rough percentage of the space sampled]
- Biggest gap: [the largest unexplored area]
Step 5: Identify Promising Regions
Flag the areas worth going deeper on.
PROMISING REGIONS:
1. [Region]: [why it's promising] — confidence: [high/medium/low]
2. [Region]: [why it's promising] — confidence: [high/medium/low]
3. [Region]: [why it's promising] — confidence: [high/medium/low]
RANKING CRITERIA: [what makes a region "promising" in this context]
Step 6: Flag Surprises and Anomalies
Note anything unexpected — these are often the most valuable findings.
SURPRISES:
- [Surprise 1]: Expected [X], found [Y] — significance: [high/medium/low]
- [Surprise 2]: Expected [X], found [Y] — significance: [high/medium/low]
ANOMALIES:
- [Anomaly 1]: [what's unusual] — worth investigating: [yes/no]
- [Anomaly 2]: [what's unusual] — worth investigating: [yes/no]
SKIP: If no surprises emerged, that itself is worth noting — the space may be well-understood.
Step 7: Exploration Summary
Synthesize findings into an actionable map.
EXPLORATION REPORT:
- Space: [one-line description]
- Key dimensions: [the 2-3 most important]
- Most promising: [top region and why]
- Biggest surprise: [most unexpected finding]
- Recommended next step: [go deep on X / explore Y / analyze Z]
Integration
Use with:
/ornt -> Orient first, then explore the identified space/anst -> Analyze the most promising regions in depth/se -> Enumerate options within a promising region/dd -> Deep-dive into a specific region after broad exploration
