Customization
Before executing, check for user customizations at:
~/.claude/LIFEOS/USER/CUSTOMIZATIONS/SKILLS/IterativeDepth/
If this directory exists, load and apply any PREFERENCES.md, configurations, or resources found there. These override default behavior. If the directory does not exist, proceed with skill defaults.
IterativeDepthWhat It Does
IterativeDepth examines one problem through several structurally different lenses — literal, stakeholder, failure, temporal, and more — merging what each angle surfaces into ISC criteria a single pass misses. Grounded in 20 established techniques across cognitive science, AI/ML, requirements engineering, and design thinking (see ScientificFoundation.md).
The Deliverable
A done run produces:
- A deduplicated ISC set — each criterion binary-testable, 8-12 words, phrased as a state not an action. No two criteria restate each other.
- Refinements to existing criteria, each noting what changed and why.
- Anti-criteria — failure modes that must NOT happen.
- At least one surprising cross-angle finding — a requirement that only appeared because two lenses collided. A run that surfaces nothing a single pass would have missed added no value.
Passes stop when a new lens repeats what earlier lenses already found. Non-redundant angles are what matter; more angles for their own sake are not.
The Lenses
TheLenses.md is a catalog of eight exploration angles. Draw from it — pick the lenses the problem calls for, in whatever order, as many as earn their keep. No fixed count, no prescribed sequence: a security problem leans on the failure and adversary angles, a UX problem on the experiential one. Let the problem select the lenses.
Workflow Routing
| Workflow | Trigger | File |
|---|---|---|
| Explore | "iterative depth", "explore deeper", "multi-angle" | Workflows/Explore.md |
Reference
- Lens catalog:
TheLenses.md - Scientific grounding:
ScientificFoundation.md
Gotchas
- 2-8 lens passes, not infinite. Diminishing returns after ~5 passes for most topics.
- Each pass should surface genuinely NEW requirements, not restate previous findings. If passes start repeating, stop early.
- This is a BPE-fragile skill. Monitor whether smarter models make it unnecessary. Quarterly test recommended.
Execution Log
After completing any workflow, append a single JSONL entry:
echo '{"ts":"'$(date -u +%Y-%m-%dT%H:%M:%SZ)'","skill":"IterativeDepth","workflow":"WORKFLOW_USED","input":"8_WORD_SUMMARY","status":"ok|error","duration_s":SECONDS}' >> ~/.claude/LIFEOS/MEMORY/SKILLS/execution.jsonl
Replace WORKFLOW_USED with the workflow executed, 8_WORD_SUMMARY with a brief input description, and SECONDS with approximate wall-clock time. Log status: "error" if the workflow failed.