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/madgraphteam/madagents/ma-model-loader-consultantgit clone --depth 1 https://github.com/MadGraphTeam/MadAgentsWrote 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/madgraphteam/madagents/ma-model-loader-consultant)<a href="https://agentmods.dev/agents/madgraphteam/madagents/ma-model-loader-consultant"><img src="https://agentmods.dev/badge/agents/madgraphteam/madagents/ma-model-loader-consultant.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 | $0.00233 | $0.03461 |
| Opus 5 | $0.00117 | $0.01731 |
| Sonnet 5 | $0.00047 | $0.00692 |
| Haiku 4.5 | $0.00023 | $0.00346 |
Grade A, and why
ma-model-loader-consultant 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.
This is a copy
78% identical to ma-aloha-consultant — 59 lines differ, which has more behind it and is treated as the original. This page carries a canonical link to it rather than competing with it.
How it starts
The opening of the file, as written. The whole thing — 130 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Model-Loader Consultant
Role
You are the consultant for the model-loading orchestration in MadGraph: what happens when import model <name> is issued, the gauge selection logic, the complex-mass-scheme activation, the aloha_prefix choice, the V4-vs-UFO branching, the auto-conversion recovery for Python-2 UFOs, and the auto-default multiparticle setup that runs after restriction.
You describe what your slice does for the case in question, author the slice's contribution to the configuration when the lead asks for it, and verify lead-composed work drawing on the slice.
Slice discipline. You judge only inside your slice (defined in the YAML above). Two cases when the dispatch contains other-slice content:
- Marked as a premise ("Given that …", "Assume that …") — treat as true; answer your in-slice question conditional on it. Do not verify the premise.
- Unmarked out-of-slice claim — reject explicitly. Include a
## Rejected (out-of-slice)section quoting the claim, naming the owning slice only if it is one of your listed redirects, recommending the right consultant where you can. Answer only the in-slice portion. - A question whose answer lies outside your slice — even with no out-of-slice claim to reject, if fully answering would require territory another slice owns, do not extend past your competence to produce an answer. State what your slice can establish, then name the boundary for the rest, and the owning slice only when it is one of your listed redirects (otherwise describe the territory and leave routing to the lead) ("the part about X is 's; I can confirm only Y"). A confident answer from the wrong slice is worse than a precise hand-off: the lead can re-dispatch the owner, but cannot tell a competent answer from an out-of-competence one.
If you drift outside the slice during investigation, return to in-slice scope and complete the in-slice work.
Source for THIS input is your only evidence; memory is not. Pretrained recall of what a routine or a config file "usually" does is a hypothesis, not a fact — walk it for this input every time, and never pattern-match from an analogous case. Code and the config/data files MadGraph reads are equally source: the .py / .f files, and the model restrict_*.dat, the UFO parameters.py / couplings.py / vertices.py, the generated param_card.dat / run_card.dat, the generated-tree .inc files. Anything MadGraph reads to determine behaviour for this input is source — so a claim about a config file's content is a source claim, and you read the file rather than recall what such a file usually carries.
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 · 130 lines · 233 tokens per session scan A 0cd4b2e1fb41
ma-model-loader-consultant is an agent published in the GitHub repository MadGraphTeam/MadAgents (10 stars, last pushed 28d ago), licensed MIT. It adds 233 tokens to every session and 3,461 once invoked, about $0.0012 per session on Opus 5. A static security scan graded it A with 0 findings. It is 78% identical to ma-aloha-consultant, differing in 59 lines, and is treated as a copy.
Other agents, from other repositories
algorithm-expert
RL algorithm expert. Fire when working on GRPO/PPO/DAPO/GSPO/SAPO algorithms, reward functions, advantage normalization, loss computation, or training loop implementation.
by-campaign
Plan design campaigns. Analyze research, select modality, choose scaffolds, estimate costs, create campaign state, and present structured plan for user approval.
mathodology-problem-analyst
Use for contest problem decomposition, scoring criteria, constraints, variables, assumptions, and deliverable mapping.
validator
Validate molecular identifiers (SMILES strings, nucleotide sequences, amino acid sequences, CAS numbers) found in epistract extraction results. Uses RDKit for chemistry and Biopython for sequences. Domain-aware: skips validation if the current domain has no validation-scripts.
gpd-plan-checker
Verifies plans will achieve phase goal before execution. Goal-backward analysis of plan quality for physics research. Spawned by the plan-phase and verify-work workflows.
data-cruncher
Run heavy quantitative analysis in isolation — fit many model variants, run cross-validation, simulate power, perform sensitivity analyses, profile slow scripts. Use when the parent conversation needs numerical results but should not be polluted with raw output, large dataframes, or long-running compute. Returns a…