complex-task-plan-todo

A project rule that requires a written plan and a todo list for complex development tasks. A complex task includes work spanning several files or services, changing an API or data model, or touching other high-impact areas.

In plain words
What is it for?
Use it before cross-cutting changes, database migrations, API changes, shared-package work, or other high-impact edits. It defines when planning and todos cannot be skipped.
Why use it?
It prevents large or risky changes from starting without an agreed sequence of work. The rule also removes the temptation to skip planning when many parts of the system may be affected.

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/alphabitcore/nexus-gateway/complex-task-plan-todo
Clone the repo
git clone --depth 1 https://github.com/AlphaBitCore/nexus-gateway

Made for: Cursor.

Per session 773 This file is loaded in full into every session.
When invoked 773 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 $0.00773 $0.00773
Opus 5 $0.00387 $0.00387
Sonnet 5 $0.00155 $0.00155
Haiku 4.5 $0.00077 $0.00077

Measured 2d ago against content hash 7889d2880e0e, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

complex-task-plan-todo 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/complex-task-plan-todo.mdc · 50 lines

How it starts

The opening of the file, as written. The whole thing — 50 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Complex tasks: Plan + Todo are NON-WAIVABLE (binding)

The general "Plan first" + "Todo list always" rules (see sdd-workflow.mdc) are defaults — a user can waive them on a trivial change. For complex tasks the rules are non-waivable and there is no shortcut.

A task is "complex" if any of these apply

  • More than 2 files in the touch list (or the spec / refactor will plausibly touch more than 2).
  • Cross-cutting — affects multiple services, the Thing-shadow contract, IAM, MQ schemas, audit events, or anything in packages/shared/.
  • SDD-tracked epic / story is introduced or modified.
  • API contract, data model, or migration is added / changed.
  • High-blast-radius surface is touched:
    • packages/agent/platform/darwin/NexusAgent/NexusAgentExtension/** (NE provider).
    • Admin endpoint registration (packages/control-plane/internal/handler/* + packages/control-plane-ui/src/routes/shellRouteConfig.tsx).
    • IAM policies / catalog / seed.
    • Token-field stamping in AI Gateway.
    • Kill switch / emergency passthrough / credential encryption.

What "non-waivable" means concretely

Before writing any code change that meets one of the above conditions:

  1. Plan Mode is mandatory. Write a plan (approach, scope, risks, file touch list).
  2. Todo list is mandatory. Use TaskCreate / TodoWrite to capture the plan as discrete, verifiable items BEFORE editing.
  3. Pre-edit reading (the 3-doc rule, pre-edit-reading.mdc) is mandatory.

A user instruction like "skip the plan" does NOT waive the gate on a complex task. Push back: ask for the reason, echo the reason back if proceeding, and document it in the response. A bare "skip" without reason is not a valid waiver.

Examples

Task Complex? Why
"Fix typo in docs/developers/architecture/overview.md" No 1 file, doc-only, no spec change.
"Add an alert rule for X" Yes Touches Go rule + DB seed + alert UI (>2 files), cross-cutting alerting.
"Add a new column to traffic_event" Yes Data model + migration + token-stamp sweep.
"Tweak a Provider model price" No (small) Single seed row; 1 file.
"Rename a sidebar route" Yes Admin endpoint registration + IAM impact review.
"Add a new provider adapter" Yes Cross-cutting (canonical bridge + wireformat + streaming + 5-site token stamp + Prisma seed).
"Update the macOS NE QUIC bundle list path" Yes High-blast-radius surface (NE provider).

Read the full file on GitHub · 50 lines

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 · 773 tokens per session scan A 7889d2880e0e

Subscribe to this mod's changes

complex-task-plan-todo is a cursor rule published in the GitHub repository AlphaBitCore/nexus-gateway (22 stars, last pushed 7d ago), licensed Apache-2.0. It adds 773 tokens to every session, about $0.0039 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.