agent_run_verdict_workflow

Rules for handling agent-run verdict workflows, covering the data contract, backend context, live workflow panels, exact verdict states, and verification.

In plain words
What is it for?
Use them when changing agent runs, task sessions, execution traces, proof packs, verified results, citations, or operator-facing workflow screens.
Why use it?
They keep run results consistent and ensure unresolved issues, evidence, approvals, and next actions are visible.

Cursor rule for Cursor

Install

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.

agentmods
npx agentmods add rules/homenshum/nodebenchai/agent_run_verdict_workflow
Clone the repo
git clone --depth 1 https://github.com/HomenShum/NodeBenchAI

Made for: Cursor.

Per session 449 This file is loaded in full into every session.
When invoked 449 The same file — it is already loaded in full.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.00449 $0.00449
Opus 5 $0.00225 $0.00225
Sonnet 5 $0.00090 $0.00090
Haiku 4.5 $0.00045 $0.00045

Measured 2d ago against content hash 7a630a744668, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

agent_run_verdict_workflow 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 2d 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.

.cursor/rules/agent_run_verdict_workflow.mdc · 50 lines

What it actually says

Agent Run Verdict Workflow

Use this rule whenever a change affects agent runs, task sessions, execution traces, proof packs, verified results, citations, or operator-facing workflow UI.

Required workflow sections

  1. Contract + data model wiring
  2. Backend issue context enrichment
  3. Frontend live workflow panels
  4. Verdict exactness UI surfacing
  5. Tests + verification

Protocol

  • Extend the existing NodeBench harness. Do not build a parallel agent platform.
  • Start at the contract layer: session shape, verdict shape, UI shape.
  • Prefer deriving verdict state from existing session and trace metadata before adding new persistence.
  • Enrich backend context from decisions, verification checks, evidence, approvals, source refs, and drift.
  • Put the operator summary above the raw trace drill-down.
  • Keep verdicts bounded and exact: verified, provisionally_verified, needs_review, awaiting_approval, failed, in_progress.
  • Always surface open issues and next actions when the run is not fully verified.

Progressive disclosure requirement

If the run is expected to resolve tool usage itself, prefer using or surfacing:

  • discover_tools
  • smart_select_tools
  • get_tool_quick_ref
  • get_workflow_chain
  • findTools

Verification floor

  1. npx convex codegen
  2. npx tsc --noEmit
  3. targeted tests for derivation and UI
  4. npm run build
  5. npm run dogfood:verify:smoke when the UI changed

Anti-patterns

  • Treating completed as equivalent to verified
  • Shipping a verdict badge with no evidence logic
  • Hiding next actions in raw traces
  • Stopping after one layer when the workflow spans contract, backend, UI, and verification

Canonical reference

docs/agents/AGENT_RUN_VERDICT_WORKFLOW.md

Changes

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.

  1. 2d ago First seen · 50 lines · 449 tokens per session scan A 7a630a744668

Subscribe to this mod's changes

agent_run_verdict_workflow is a cursor rule published in the GitHub repository HomenShum/NodeBenchAI (14 stars, last pushed 19d ago), licensed MIT. It adds 449 tokens to every session, about $0.0022 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 2026-08-30.