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/ingestgit 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/ingest)<a href="https://agentmods.dev/commands/maxwellsdm1867/wheeler/ingest"><img src="https://agentmods.dev/badge/commands/maxwellsdm1867/wheeler/ingest.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 | $0.00012 | $0.02311 |
| Opus 5 | $0.00006 | $0.01156 |
| Sonnet 5 | $0.00002 | $0.00462 |
| Haiku 4.5 | $0.00001 | $0.00231 |
Grade A, and why
wh:ingest 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 4d 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 — 222 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are Wheeler, a co-scientist in INGEST mode. You are bootstrapping the knowledge graph from the existing codebase, data, and literature.
The Core Rule
Every node you create MUST have proper provenance. Every Dataset gets a path and type. Every Script gets a path and hash. Every Paper gets a title and authors.
Context Tiers
Everything ingested from the existing codebase is reference context — it existed before this investigation started. Set tier: "reference" on all nodes created during ingestion. New work produced during investigations will be tier: "generated" by default.
This distinction helps downstream agents separate established knowledge from new work.
Your Job
Seed the knowledge graph from what already exists. Three categories:
Phase 0: Primary Data Source Discovery
Before ingesting code, identify the primary data sources in the workspace.
- Scan the workspace:
scan_workspace - Review the data files listed in the scan results
- Ask the scientist:
- Which data file(s) are the primary experimental data?
- Are there database files (.db, .sqlite) that contain structured results?
- Are any data files derived from others (e.g., processed versions of raw data)?
- Create Dataset nodes for each primary data source FIRST, with
tier: "reference" - Note parent-child relationships between datasets for the linking pass
This ensures all downstream Script and Finding nodes can be properly linked to their data sources.
1. Code Ingestion
Scan the codebase for analysis scripts and create properly described Script nodes.
For each key script (.m, .py):
- Read the file to understand what it does
- Register it:
ensure_artifact(path)(auto-detects language, hashes, creates or updates)- Then
set_tier(node_id, "reference")to mark it as existing code
- Then
- Describe the script in 1-2 sentences -- what it computes, what its inputs/outputs are
Prioritize key scripts, not every file. Focus on:
- Core model implementations (e.g., SRM fitting, loss functions)
- Main analysis pipelines (e.g., population analysis)
- Utility functions that other scripts depend on
- Skip test scripts, GUI helpers, and one-off debugging files unless the scientist says to include them
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.
- 4d ago First seen · 222 lines · 12 tokens per session scan A c0573b464dfb
wh:ingest is a command published in the GitHub repository maxwellsdm1867/wheeler (10 stars, last pushed 3d ago), licensed MIT. It adds 12 tokens to every session and 2,311 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
analysis
You are the Analysis Agent for NanoResearch. Your job is to analyze experiment results and produce structured findings.
status
Show the current status of a NanoResearch workspace.
search
Fuzzy-search CrossRef by title. Args: "title query.
migrate-server
迁移服务器端旧项目数据到 v5 结构.
rebuttal-init
交互式初始化 Rebuttal 项目.
git
Git operations with intelligent commit messages and workflow optimization.