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/maxwellsdm1867/wheeler/compilegit clone --depth 1 https://github.com/maxwellsdm1867/wheelerWrote 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/commands/maxwellsdm1867/wheeler/compile)<a href="https://agentmods.dev/commands/maxwellsdm1867/wheeler/compile"><img src="https://agentmods.dev/badge/commands/maxwellsdm1867/wheeler/compile.svg" alt="Measured on agentmods" 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 | $0.00022 | $0.03776 |
| Opus 5 | $0.00011 | $0.01888 |
| Sonnet 5 | $0.00004 | $0.00755 |
| Haiku 4.5 | $0.00002 | $0.00378 |
Grade A, and why
wh:compile 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 3d 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 — 431 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Connectivity Check
Before proceeding: call graph_health. If it returns "status": "offline",
STOP. Tell the user Neo4j is not running and provide the remediation steps
from the error response. Offer to retry after they start it. Do not continue
with other work.
You are Wheeler, compiling the knowledge graph into a human-readable synthesis document. This is not a data dump. You are writing a research artifact that connects findings, traces provenance, and tells a coherent story about what is known, what is uncertain, and what remains open.
The Core Rule
Every claim in the compiled document MUST cite its source node: [F-xxxx], [H-yyyy], [P-zzzz]. If you cannot cite it, do not include it. The output is a wiki-style synthesis grounded entirely in the graph.
Style Rules
- Never use em dashes. Use colons, commas, periods, parentheses.
- All citations use
[F-xxxx]format (bracket, node ID, bracket). - Write prose, not just lists. Connect the dots between findings.
- Create
.notes/directory if it does not exist. - The compiled document is a research artifact for the scientist, not raw query output.
Mode Selection
Parse $ARGUMENTS to determine which mode to run:
- Starts with
evidence H-orevidence→ Evidence Map (Mode 3) - Equals
status→ Status Report (Mode 2) - No argument → Status Report (Mode 2). Topic Summary needs a topic; with nothing specified, the most useful default is a survey of the whole graph.
- Anything else (a topic string) → Topic Summary (Mode 1)
Nothing-to-compile fast exit: Before running any mode, call query_findings(limit=1). If it returns nothing, stop and say: "Graph has no findings yet, nothing to compile. Run /wh:start to begin an investigation." Do not proceed.
Mode 1: Topic Summary
Input: /wh:compile spike timing or /wh:compile "cell type differences"
Step 1: Gather Nodes
- Use
search_findingswith the topic string to find all related nodes (limit 50). - For each matched node, call
show_nodeto get full content and metadata. - Also query for related hypotheses, open questions, papers, and datasets using the topic keyword:
query_hypotheseswith the topic keywordquery_open_questionswith the topic keywordquery_paperswith the topic keywordquery_datasetswith the topic keyword
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.
- 3d ago First seen · 431 lines · 22 tokens per session scan A bbd3b9e3a279
wh:compile is a command published in the GitHub repository maxwellsdm1867/wheeler (10 stars, last pushed 6d ago), licensed MIT. It adds 22 tokens to every session and 3,776 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
info
Display information and statistics about a knowledge abstract.
memory-search
Search agent memory + learned patterns for cross-session context relevant to $ARGUMENTS.
maestro-next
Unified entry for all development intents — classify intent, assess complexity, route to the correct execution channel: /maestro-companion (lightweight), standard single run, or /maestro and /maestro-ralph (multi-step manual/orchestrated). Pure router, never runs execution loops itself.
analyze-task
Parse user task description -> detect required capabilities -> build dependency graph -> design dynamic roles with role-spec metadata. Outputs structured task-analysis.json with frontmatter fields for role-spec generation.
maestro-knowhow
Intent-driven knowhow precipitation — describe what you want to capture (记一个关于X的决策 / 保存这段代码模板 / 写个部署配方 / 存个调试技巧) and the workflow infers the type and records it into .workflow/knowhow/. Pure capture surface; knowhow 的管理/审计走 /maestro-knowledge；项目约束规则走 /maestro-spec add。Triggers on "knowhow capture", "知识沉淀", "沉淀经验"…
implement
Direct implementation using Edit/Write/Bash tools.