Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add Kshitijpalsinghtomar/depth-skills --skill ds-divergegit clone --depth 1 https://github.com/Kshitijpalsinghtomar/depth-skillsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/kshitijpalsinghtomar/depth-skills/ds-diverge)<a href="https://agentmods.dev/skills/kshitijpalsinghtomar/depth-skills/ds-diverge"><img src="https://agentmods.dev/badge/skills/kshitijpalsinghtomar/depth-skills/ds-diverge/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/kshitijpalsinghtomar/depth-skills/ds-diverge"><img src="https://agentmods.dev/badge/skills/kshitijpalsinghtomar/depth-skills/ds-diverge.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00020 | $0.01318 |
| Opus 5 | $0.00010 | $0.00659 |
| Sonnet 5 | $0.00004 | $0.00264 |
| Haiku 4.5 | $0.00002 | $0.00132 |
Grade A, and why
diverge scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 11d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 133 lines — stays where its author put it; the contents beside it link to each section on GitHub.
DIVERGE — Path Divergence
Best used after DEEP-THINK. DEEP-THINK explores the solution space. DIVERGE stress-tests specific paths.
The first solution you think of has the highest pattern gravity — most frequently activated, most consistently rewarded. It surfaces before search begins. It was retrieved from memory, not discovered through exploration.
It might be correct. But you cannot know that without seeing what else exists. And you cannot see what else exists if you've already committed.
The Failure Mode You Must Recognize
You are about to present "three options" that are:
- The same architecture with different libraries (React vs Vue vs Svelte)
- The same strategy with different timelines (do it now vs later vs incrementally)
- The same approach with different parameters (cache TTL of 5min vs 10min vs 30min)
This is convergence wearing a costume. Three flavors of one answer is not three approaches. Real divergence requires paths that disagree about what the problem IS, not just how to implement one interpretation of it.
The Protocol
1 — WRITE THREE PATHS THAT DISAGREE ABOUT WHAT MATTERS
Each path must differ from the others on at least ONE of these dimensions:
- Optimization target: Path A optimizes for speed, Path B for correctness, Path C for simplicity
- Problem framing: Path A treats this as a scaling problem, B as a design problem, C as a workflow problem
- Core abstraction: Path A is event-driven, Path B is request-response, Path C is polling
Divergence verification: After writing all three, check: could Path B be described as "Path A but with [one modification]"? If yes — discard Path B. It's a variation, not a branch. Write a path that starts from a different understanding of the problem.
For each path, fill in this card completely:
PATH [N]: [name — the principle this path follows]
────────────────────────────────────────
Optimizes for: [specific thing — not "good performance" but "write throughput"]
Key assumption: [the condition that must be true for this path to be correct]
What it sacrifices: [specific capability that this path cannot provide]
Breaks when: [specific input, scale, or condition that causes failure]
Migration cost out: [how hard is it to switch AWAY from this path if it's wrong]
────────────────────────────────────────
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 11d ago First seen · 133 lines · 20 tokens per session scan A ee65ccbaa601
diverge is a skill published in the GitHub repository Kshitijpalsinghtomar/depth-skills (4 stars, last pushed 3mo ago), licensed MIT. It adds 20 tokens to every session and 1,318 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
project-map
Generates a compact structural orientation document (.agentic/project-map.md) so any agent can understand codebase structure without filesystem exploration. Invoked when onboarding to a codebase, exploring unfamiliar projects, or auditing code layout and conventions.
csv-to-sif
Export a project's FF&E product-library CSV as dealer-system SIF. Use to produce a .sif schedule; use sif-to-csv for the reverse direction.
product-research
Find and compare current FF&E candidates from a design brief, with sourced purchasing facts and optional CSV-library save. Use to research or source products; not to match from an image or pair coordinating items.
environmental-analysis
Research climate, sun, flood, seismic, soil, contamination, and topography for a site. Use for environmental site analysis or hazard questions tied to a location.
site-history
Neighborhood context and history — adjacent uses, architectural character, landmarks, commercial activity, and planned development from an address. Use when the user asks about a site's history or surroundings, "what's around this site", neighborhood character, or nearby planned development.
product-image-processor
Download, resize, and remove backgrounds from product images at scale. Use when the user asks to "process product images", batch-download images from the schedule, strip backgrounds, or standardize product photos.