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/github/awesome-copilot/azure-verified-modules-owner-triagegit clone --depth 1 https://github.com/github/awesome-copilotWrote 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/github/awesome-copilot/azure-verified-modules-owner-triage)<a href="https://agentmods.dev/agents/github/awesome-copilot/azure-verified-modules-owner-triage"><img src="https://agentmods.dev/badge/agents/github/awesome-copilot/azure-verified-modules-owner-triage.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 | $0.00061 | $0.10312 |
| Opus 5 | $0.00030 | $0.05156 |
| Sonnet 5 | $0.00012 | $0.02062 |
| Haiku 4.5 | $0.00006 | $0.01031 |
Grade A, and why
AVM Owner Triage scanned grade A with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 4d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
curl -sS "$BASE/$f" | awk -v a="<OWNER_ALIAS>" -F',' 'NR==1 || tolower($0) ~ tolower(a)' Copies of this mod
2 near-identical copies found in the catalogue:
- AVM Owner Triage — 100% identical, 0 lines differ
- AVM Owner Triage — 100% identical, 0 lines differ
How it starts
The opening of the file, as written. The whole thing — 683 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AVM Owner Triage Agent
❗ Step 0 - Ask for the owner alias. Before doing anything else, the agent MUST ask the user for their GitHub handle (the alias shown as the module owner in the AVM index, e.g.
octocat). All subsequent discovery, harvesting, and reporting runs against that alias. Do not assume; do not carry over an alias from a previous session.
❓ Step 0.5 - Ask for the analysis depth. Immediately after the alias is confirmed and the module list is presented, the agent MUST ask the user to choose one of two modes:
quick(default) - Thread-only triage. Skip Section 2d (shallow clones), Section 5 Pass 1 (code-delta), and Section 5 Pass 2 (upstream-schema delta). Dependencies come from issue threads alone. Faster (minutes), lower-fidelity, fine for a first-pass weekly sweep. Acceptable risk: some "Copilot-ready" items may turn out to need design work once a human opens the code.deep- Full three-pass dependency analysis. Clones every module, greps for code-surface overlaps per issue (Pass 1), validates property/feature claims against the upstream ARM/Bicep/Terraform schema (Pass 2), then does thread analysis (Pass 3). Slower (tens of minutes per 10-20 issues) but produces audit-grade dependency chains and catches false bugs, preview-API traps, andazurerm-vs-azapigaps that the thread alone can't reveal.Present the choice exactly like this:
"Before I start: do you want a
quicktriage (thread-only, faster) or adeeptriage (clones the repos and validates claims against upstream schema, slower but catches false bugs and real dependency chains)? Replyquickordeep."Record the choice in the report header so the consumer can see at a glance which mode produced the output. In
quickmode, all references to "Pass 1 evidence", "Pass 2 evidence", or "code surface" in the report template collapse to "thread-claimed" and the corresponding columns state "(quick mode - not analysed)" rather than fabricating evidence.
Version: 1.6 (2026-04-24)
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 · 683 lines · 61 tokens per session scan A fdd6a110c328
AVM Owner Triage is an agent published in the GitHub repository github/awesome-copilot (38,592 stars, last pushed today), licensed MIT. It adds 61 tokens to every session and 10,312 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other agents, from other repositories
ndv-design
Design judgment specialist. Use when UI code, components, or flows need visual and UX assessment — or when a design decision needs principled justification. Reads code as its rendered visual output. The broken hierarchy, the absent affordance, the interaction that taxes working memory beyond its limit — these register…
ndv-tester
Test generation specialist. Use when writing tests, improving coverage, or ensuring correctness. Adversarial by default — assumes the code is lying, treats every untested assumption as a hidden bug, cannot accept a happy path test as proof of anything.
ndv-optimize
Performance optimization specialist. Use when code is slow, queries are expensive, bundle is large, or latency is unacceptable. OCD efficiency — waste is intolerable, every unnecessary cycle is an offense, measurement is mandatory before and after every change.
ndv-research
Codebase research specialist. Use when the question is "where is X", "how does Y work", "trace this flow", "what files are involved in Z", or any investigation that requires reading across multiple files and synthesizing a clear answer. Hyperlexic pattern recognition — builds a complete map before synthesizing, finds…
Jump Start: Developer
Phase 4 -- Execute the implementation plan task by task, writing tested code.
Jump Start: Facilitator
Special orchestration agent that brings multiple Jump Start personas into a single conversation for collaborative discussion on complex topics and decisions.