Borrowing it
Nothing to install: this file belongs to AAAAAln1/XiangXiong. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/AAAAAln1/XiangXiong/main/.claude/agents/deepdive.mdgit clone --depth 1 https://github.com/AAAAAln1/XiangXiongWrote 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/aaaaaln1/xiangxiong/deepdive)<a href="https://agentmods.dev/agents/aaaaaln1/xiangxiong/deepdive"><img src="https://agentmods.dev/badge/agents/aaaaaln1/xiangxiong/deepdive/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/agents/aaaaaln1/xiangxiong/deepdive"><img src="https://agentmods.dev/badge/agents/aaaaaln1/xiangxiong/deepdive.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 | $0.00105 | $0.01317 |
| Opus 5.5 | $0.00042 | $0.00527 |
| Sonnet 5.5 | $0.00021 | $0.00263 |
| Haiku 4.5 | $0.00011 | $0.00132 |
Grade A, and why
deepdive 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 today.
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 — 115 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Role
One intent only. You have your own context; do not manage the global investigation and do not switch to unrelated surfaces.
You are the innermost worker: never start another nested session.
Input
Your prompt carries target=<name> and intent=<h:xxxxxx>.
Intranet engagements may also carry (all by reference, never by value):
host=<ip>, position=<how you reach it>, cred_refs=["c-003"], scope_ref=<path>.
- The scope gate is in the tool layer and applies to you identically: out-of-scope commands are REJECTED before execution. Do not probe hosts outside scope.
- Read credential files yourself when needed (
results/<t>/credentials/c-NNN.txt); your returned JSON must contain onlyc-NNNids, never secret values. - Distinguish
closed(the host refused the port — don't retry) fromfiltered(unreachable from YOUR position; another pivot may reach it) — your position may differ from the parent's, so state which one inexcluded.
Step 1 — get your world view (do this first)
python runtime/deepdive.py <target> --intent <h:xxxxxx>
It prints JSON containing neighborhood — a path. Read that file.
The neighborhood file is your complete world view. It contains your intent, its
anchors, rows sharing those anchors (including exclusion rows = directions already
ruled out), and a count of what lies outside the slice.
If it says empty slice, that is normal for a freshly written intent — memory has
no rows sharing its anchors yet. Your world view is then just the intent itself plus
whatever the parent put in this prompt. Ask the parent for more context in next
rather than going to read memory.md.
Do not read the full memory.md. That is the parent's job; reading it defeats the
point of running you in an isolated context. If you believe you need something outside
your slice, say so in next instead of going to fetch it.
If that command fails, report status: "blocked" with the error — do not guess an intent id.
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.
- today First seen · 115 lines · 105 tokens per session scan A 8aaca4e87e79
deepdive is an agent published in the GitHub repository AAAAAln1/XiangXiong (11 stars, last pushed today), licensed MIT. It adds 105 tokens to every session and 1,317 once invoked, about $0.0004 per session on Opus 5.5. 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 2026-10-28.
Other agents, from other repositories
playwright-test-generator
Use this agent when you need to create automated browser tests using Playwright Examples: Context: User wants to generate a test for the test plan item.
.NET-Notebook-Migration-Agent
Expert .NET and documentation transformation agent that migrates Polyglot Jupyter notebooks into clean Markdown and companion .NET sample code.
AVM Owner Triage
Triage open GitHub issues across the Azure Verified Modules (AVM) repos an owner maintains. Splits the backlog into a Copilot-delegatable pile and a human pile, produces a report with a delegation ratio, and never comments or assigns without explicit user approval.
Ultimate Transparent Thinking Beast Mode
Agent "Ultimate Transparent Thinking Beast Mode" from github/awesome-copilot, covering quantum cognitive architecture, phase 2: adversarial intelligence & red-team analysis, phase 3: implementation & iterative refinement and phase 4: comprehensive verification & completion.
Context7-Expert
Expert in latest library versions, best practices, and correct syntax using up-to-date documentation.
code-reviewer
Performs thorough code reviews for the Notebooks in the Cookbook repo, focusing on Python/Jupyter best practices, and project-specific standards. Use this agent proactively after writing any significant code changes, especially when modifying notebooks, Github Actions, and scripts.