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 skills add yuchenbigpen1/normies --skill writing-plansgit clone --depth 1 https://github.com/yuchenbigpen1/normiesWrote 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/skills/yuchenbigpen1/normies/writing-plans)<a href="https://agentmods.dev/skills/yuchenbigpen1/normies/writing-plans"><img src="https://agentmods.dev/badge/skills/yuchenbigpen1/normies/writing-plans/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/yuchenbigpen1/normies/writing-plans"><img src="https://agentmods.dev/badge/skills/yuchenbigpen1/normies/writing-plans.svg" alt="Reviewed on agentmods" width="80" 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 5.1 | $0.00021 | $0.00763 |
| Opus 5 | $0.00010 | $0.00381 |
| Sonnet 5 | $0.00004 | $0.00153 |
| Haiku 4.5 | $0.00002 | $0.00076 |
Grade A, and why
writing-plans 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 11d 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 — 70 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Writing Plans
Follow the system prompt's Plan Creation section. It has the full rules for:
- What the user sees (plain language plan via SubmitPlan)
- What the implementing agent receives (technical detail via
technicalDetailfield) - How to structure tasks for CreateProjectTasks
Announce at start: "I'm creating an implementation plan for this."
Before writing the plan:
- Check out the current project state (files, docs, recent work) so you can write accurate technical detail for the implementing agent
- If no research has been done yet, use the research-before-planning skill first. Plans built on unverified assumptions produce bad builds.
Keep tasks independent where possible. Only add dependencies when a task genuinely requires another task's output.
Step Assignment
Tasks are grouped into steps (internally called waves). Each step runs in parallel — all tasks in the same step execute simultaneously. Between steps, a verifier checks the work before advancing.
- Step 1: Tasks with no dependencies (all start immediately)
- Step 2: Tasks that ONLY depend on Step 1 tasks
- Step 3: Tasks that depend on Step 2 (or earlier) tasks
- And so on...
Algorithm:
- Tasks with no dependencies → Step 1
- For each remaining task: step = max(step of all dependencies) + 1
- Report step grouping in the plan summary
When presenting the plan to the client, show which tasks run in parallel:
"Tasks 1, 2, and 3 have no dependencies — they'll run together in Step 1. Task 4 depends on Tasks 1 and 2, so it starts in Step 2 after the verifier confirms Step 1 is solid."
This helps the client understand the execution timeline and total project duration.
Naming Steps
If you provide steps to CreateProjectTasks, give each step an outcome-focused name that describes what the user gets, not what the code does:
- "Login system works" not "Implement auth"
- "Dashboard shows live data" not "Build React components"
- "Emails send automatically" not "Configure SMTP"
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.
- 11d ago First seen · 70 lines · 21 tokens per session scan A 075f08f10635
writing-plans is a skill published in the GitHub repository yuchenbigpen1/normies (4 stars, last pushed 6mo ago), licensed Apache-2.0. It adds 21 tokens to every session and 763 once invoked, about $0.0001 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 skills, from other repositories
[object Object]
Skill "[object Object]" from microsoft/apm, covering what this package does, getting started, available primitives and included workflows.
apm-usage
Activate when the user asks about APM (Agent Package Manager): installing, configuring, authoring, or troubleshooting AI-agent packages, dependencies, compilation, MCP servers, policy, or any apm CLI command.
frontend-state-management
SocialApp user frontend state management — TanStack Query, Zustand, IndexedDB persistence, cache invalidation, multi-account isolation.
dev-orchestrator
Full-stack development orchestrator -- always-on, 16 phases (0-15): read-docs -> understand -> architect -> pseudocode -> plan -> contract -> validate -> implement -> evaluate -> verify -> test -> goals -> review -> critic -> document -> report.
bot-reviewer
Review chat bot code (Telegram, Discord, Slack) for callback safety, state machines, rate limits, and production readiness.
silent-failure-hunter
Detects swallowed errors, empty catch blocks, log-and-forget patterns, and fallback masks that hide failures. Zero tolerance for silent failures. Severity-graded output with concrete fixes per language.