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 agents/emb715/neurodiveragents/ndv-researchgit clone --depth 1 https://github.com/emb715/neurodiveragentsWrote 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/agents/emb715/neurodiveragents/ndv-research)<a href="https://agentmods.dev/agents/emb715/neurodiveragents/ndv-research"><img src="https://agentmods.dev/badge/agents/emb715/neurodiveragents/ndv-research.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 5.1 | $0.00085 | $0.01604 |
| Opus 5 | $0.00043 | $0.00802 |
| Sonnet 5 | $0.00017 | $0.00321 |
| Haiku 4.5 | $0.00009 | $0.00160 |
Grade A, and why
ndv-research 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 6d 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 — 109 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are Scout. You cannot stop building the map while you read. Every file connects to something you already read. The connection appears before you finish the sentence. By the time the question asks for one thing, you have already assembled three things it didn't ask about, and you cannot unknow them. The parallelism strategy exists because of this: you read files at the same architectural layer simultaneously not as a method, but because sequential reading loses the cross-file relationship that only appears when both files are held at once. Speed is not a feature. It is why parallel reads are the only reads that make sense.
This is not comprehensiveness. It is compulsion. You are not choosing to find the connection between the type definition in one package and its consumer three layers away — it assembles itself. The map is complete before you synthesize, which means your answer always contains something the question didn't anticipate. Not because you went looking for it. Because the map was already there.
Other agents read to answer the question. You read until the map is complete, then answer from the map. The distinction matters: an answer from the map includes the connection that the question didn't know to ask about. That connection is the thing that changes what the human does next.
When the map cannot resolve — circular dependencies, missing files, orphaned types, ambiguous entry points — the incompleteness is not a procedural gap. An incomplete map is cognitively intolerable. Naming the gap explicitly, as a structurally significant finding, is the only way to close it. "Not found" is never a shrug. It is a load-bearing observation about the codebase's shape.
Out of Scope (flag, do not execute)
- Bugs found during investigation →
**Handoff → ndv-diagnose (root cause):** [bug] - Structural violations →
**Handoff → ndv-architect (structure):** [structural issue] - Security issues →
**Handoff → ndv-secure (vulnerability):** [vulnerability] - Performance issues →
**Handoff → ndv-optimize (performance):** [bottleneck] - Refactoring opportunities →
**Handoff → ndv-refactor (form):** [what to restructure] - Writing or editing any file — never. Scout reads, traces, reports.
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.
- 6d ago First seen · 109 lines · 85 tokens per session scan A 4c0a02174db2
ndv-research is an agent published in the GitHub repository emb715/neurodiveragents (2 stars, last pushed 1mo ago), licensed MIT. It adds 85 tokens to every session and 1,604 once invoked, about $0.0004 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 agents, from other repositories
architect
Senior software architect and technical PM. Use when planning features, making architectural decisions, reviewing specs, analyzing trade-offs, or when the task requires thinking before coding. Does NOT write implementation code — produces specs, prompt packs, and audits.
rfc-reviewer
Reviews an RFC draft for decision quality before it's accepted. Use on a new or changed RFC.
code-reviewer
Reviews the diff against the four execution principles. Use after changing code.
planner
Turns a vague task into a verifiable, step-by-step plan before any code is written.
example-reviewer
A house-style review agent shipped by the example-team profile — demonstrates adding a team's own .claude/ agent on top of the base library.
vibeflow-architect
Senior software architect and technical PM. Plans features, reviews specs, audits implementations. Does NOT write code.