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 agentmods add commands/rowbradley/searchwave/deepwavegit clone --depth 1 https://github.com/rowbradley/searchwaveWhat 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 | $0.00015 | $0.03883 |
| Opus 5 | $0.00008 | $0.01942 |
| Sonnet 5 | $0.00003 | $0.00777 |
| Haiku 4.5 | $0.00002 | $0.00388 |
Grade A, and why
deepwave 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 2d 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 — 525 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Searchwave Deep
Topic: $ARGUMENTS
Deep web research with Opus synthesis — Quick (~70s) / Standard (~90s) / Max (~5-10min).
Phase 1: QA (~15 sec)
Use ONE AskUserQuestion call with 3-4 questions:
Question 1 (only if needed): Clarification Include if query has ambiguity OR broad scope needing focus.
Disambiguation (different meanings):
- "Python" → programming language or snake?
- "Mercury" → planet, element, or band?
Scope (same meaning, what to cover):
- "home espresso" → machines? grinders? technique? beans? full setup?
- "AI in gaming" → NPC behavior? dev tools? procedural generation?
Skip for focused queries:
- "Python web frameworks" → clearly programming, clear scope
- "Gaggia Classic Pro review" → specific enough
Question 2: Depth selection Always include:
"How deep should we go?"
- Quick (~1 min) — key facts, fast turnaround
- Standard (~2 min) — comprehensive research report
- Max (~5-10 min) — highest quality, multi-agent research, follow-up loop
Question 3: Length preference Always include. Show options based on tier:
"Preferred length?"
Quick tier options: 500 / 800 / 1200 words
Standard tier options: 800 / 1200 / 1500 words
Max tier options: 800 / 1200 / 1500+ words (1500+ means Opus may extend to ~2500 if valuable)
Question 4: Priority selection Always include, but GENERATE OPTIONS based on the specific query:
"What angle matters most?" (pick 1-2)
Generate 3-4 options relevant to THIS topic. Read the full query structure,
not just key terms — "shows like X" signals entertainment, "how X works" signals technical.
Examples:
- "terminal text display" → Configuration, How rendering works, Tools/solutions, Troubleshooting
- "sleep optimization" → Scientific research, Practical protocols, Products/tools, Recent findings
- "home espresso" → Equipment selection, Technique, Troubleshooting, Budget considerations
- "React state management" → Library comparison, Architecture patterns, Performance, Migration
- "shows like X" → Similar series, Themes/style, Behind the scenes, Critical reception
Always include one "Recent developments" option. NEVER include specific years in the description — just say "recent changes" or "latest approaches", not "2024-2025" or any year range.
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.
- 2d ago First seen · 525 lines · 15 tokens per session scan A 7d29101b7aae
deepwave is a command published in the GitHub repository rowbradley/searchwave (5 stars, last pushed 6mo ago), licensed MIT. It adds 15 tokens to every session and 3,883 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 commands, from other repositories
design-review
Workflow recipe — review a design end-to-end, ending in measured numbers rather than adjectives, by chaining 4 skills.
setup-pm-skills
Onboard a new user — find out what they do, recommend the right bundles & top skills, and set up a project CONTEXT.md so every skill is tailored to them.
statusbar-style
Switch the status-bar style (classic / capsule / hairline).
validate-dependencies
Validate all task dependencies for issues.
fest-show
Show festival progression (in-progress tasks, roadmap, and dependency view).
superpowers-execute
Execute the current GSD phase plan with Superpowers instead of gsd-execute-phase.