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 instructions/misterbrookt/finreportbench/agents-mdgit clone --depth 1 https://github.com/MisterBrookT/finreportbenchWrote 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/instructions/misterbrookt/finreportbench/agents-md)<a href="https://agentmods.dev/instructions/misterbrookt/finreportbench/agents-md"><img src="https://agentmods.dev/badge/instructions/misterbrookt/finreportbench/agents-md.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 | $0.02610 | $0.02610 |
| Opus 5 | $0.01305 | $0.01305 |
| Sonnet 5 | $0.00522 | $0.00522 |
| Haiku 4.5 | $0.00261 | $0.00261 |
Grade A, and why
finreportbench AGENTS.md 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 — 309 lines — stays where its author put it; the contents beside it link to each section on GitHub.
FinReportBench Agent Handoff
This file is the operational contract for maintainers and coding agents. Read it before changing data, prompts, evaluation logic, or frozen experiment assets. Only repository-verifiable facts belong here.
Mission
FinReportBench measures whether generated financial reports are suitable for institutional delivery. The active release contains:
- 244 bilingual flash-report tasks: 122 Chinese and 122 English;
- one scripted research trajectory per task;
- one frozen source bundle per task;
- a 35-item hierarchical evaluator: G0 deliverability, G1 report identity, and G2 institutional completeness;
- frozen nine-model benchmark, evaluator-calibration, and five-model skill-evolution evidence;
- the submitted paper, supplementary material, and checklist.
The active benchmark is semantic_v1. final_v1 and Flash60 are retained as
historical calibration assets, not as the public dataset entry point.
Start Here
| Need | Canonical path |
|---|---|
| Project status | docs/status.md |
| Reproduction guide | docs/reproduction/README.md |
| Release validation | scripts/validate_release.py |
| Public task package | benchmark/l3_generation_packages/semantic_v1_flash244_scripted_trajectory_20260728/sample_244.jsonl |
| Query cards and hidden provenance | benchmark/queries/semantic_v1/ |
| Frozen source bundle | benchmark/source_bundle/semantic_v1/ |
| Source-bundle manifest | benchmark/source_bundle/semantic_v1/manifest.json |
| Source-bundle builder | scripts/build_semantic_v1_source_bundle.py |
| Active rubric | benchmark/rubrics/g0_itemized.json |
| Judge workflow | docs/annotation/g0_judge_workflow.md |
| Canonical nine-model result | runs/g0_judge_eval/20260729_semantic_v1_flash244_partner_completed_9model_luna/ |
| Evaluator calibration | runs/evaluator_calibration/20260729_5case_3evaluator/ |
| Skill evolution | runs/skill_evolution/20260729_30case_5model_claude_sdk_hybrid/ |
| Paper snapshot | publication/aaai27/ |
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 · 309 lines · 2,610 tokens per session scan A 86441f11c08d
finreportbench AGENTS.md is an instructions file published in the GitHub repository MisterBrookT/finreportbench (2 stars, last pushed 27d ago), licensed MIT. It adds 2,610 tokens to every session, about $0.0130 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 instructions, from other repositories
agentic-vbench AGENTS.md
AGENTS.md instructions for PhiloLabs/agentic-vbench, covering agents.md, start, map, docs and code.
benchjack CLAUDE.md
Instructions for benchjack/benchjack, covering claude.md, project overview, tech stack, repository layout and common commands.
mcpscope AGENTS.md
Instructions for ffleurey/mcpscope, covering mcpscope agent guide, fast start, project shape, working style and inspect a running mcpscope with mcpscope.
agentic-vbench CLAUDE.md
Claude Code instructions for PhiloLabs/agentic-vbench, a project described as: AgenticVBench: Can AI Agents Complete Real-World Post-Production Tasks?
Tracely-ai CLAUDE.md
Claude Code instructions for Jwuthri/Tracely-ai, covering claude.md, commands, architecture, hard rules and gotchas.
mobilegym AGENTS.md
Instructions for Purewhiter/mobilegym, covering agents.md, project overview, type-checking strategy, eslint and with data graph generation.