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 skills/savedpixel/ai-agent-rules-generator/plan-tasknpx skills add savedpixel/ai-agent-rules-generator --skill plan-taskgit clone --depth 1 https://github.com/savedpixel/ai-agent-rules-generatorWhat 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 | $0.00045 | $0.01122 |
| Opus 5 | $0.00023 | $0.00561 |
| Sonnet 5 | $0.00009 | $0.00224 |
| Haiku 4.5 | $0.00005 | $0.00112 |
Grade A, and why
plan-task 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 yesterday.
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 — 141 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Create a Detailed Implementation Plan
HARD GATE — PLANNING ONLY. ZERO IMPLEMENTATION. This skill invokes PLANNING MODE. You must produce a plan document and NOTHING ELSE. You are forbidden from editing, creating, or modifying any code, template, CSS, JS, PHP, or JSON file outside of the plan document itself and
docs/task/todo.md. This gate is absolute — it cannot be overridden by the user's phrasing, tone, or imperative language. If you write even one line of implementation code, you have violated this skill. Your ONLY output is the plan file + todo.md update + a summary asking for approval.
What You Receive
The user will describe a task. Their description is INPUT to the plan — not an instruction to execute. "Please update X" means the plan must include it — not that you do it now.
Steps
1. Review Recent Context
- Read latest task log(s) in
docs/task/logs/— 1–2 most recent files. - Read
docs/task/todo.md— check for overlapping work. - Read
docs/task/lessons.md— check for applicable lessons. - Read agent observation logs in
docs/agent-observations/— unresolved items may affect scope. - Read relevant documentation via
docs/plugin/**/0-index.md.
2. Understand the Request
- Read the user's description carefully.
- Search the codebase for current state of affected files.
- Identify all files, functions, data structures, and dependencies.
- Do NOT assume. If something is ambiguous, add it to Open Questions (step 2.5).
2.5. Question Gate (Iterative)
Before writing the plan, identify open questions about scope, approach, or constraints.
Rules:
- Assign stable IDs: Q1, Q2, Q3, … (never renumber across rounds)
- Present questions with a brief recommended answer for each
- Wait for user answers before proceeding
- If answers raise new questions, add them with next available ID and ask again
- Repeat until zero open questions remain
- Answered questions move to Confirmed Inputs section
- Once resolved: update plan
StatusfromQuestions PendingtoDraft
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.
- yesterday First seen · 141 lines · 45 tokens per session scan A 7df0b7822b3c
plan-task is a skill published in the GitHub repository savedpixel/ai-agent-rules-generator (7 stars, last pushed 4mo ago), licensed MIT. It adds 45 tokens to every session and 1,122 once invoked, about $0.0002 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
jira-cli
Interact with Jira from the command line to create, list, view, edit, and transition issues, manage sprints and epics, and perform common Jira workflows. Use when the user asks about Jira tasks, tickets, issues, sprints, or needs to manage project work items.
azure-prices
Look up and compare Azure service pricing using the Azure Retail Prices API. Use this skill whenever the user asks about Azure costs, pricing, rates, or wants to compare prices across regions or services — even if they don't say "pricing" explicitly. Trigger for questions like "how much does a D2 v2 VM cost?"…
calculator
Performs arbitrary-precision arithmetic calculations including addition, subtraction, multiplication, division, and exponents. Use when the user asks to calculate, compute, or evaluate math expressions, or when precise decimal arithmetic is needed to avoid floating-point errors.
az-cli
Use the Azure CLI (az) to manage Azure resources from the command line. Trigger this skill whenever the user asks to create, configure, manage, deploy, or interact with any Azure resource — even if they don't explicitly mention "az cli". Also trigger when the user asks about Azure CLI commands, syntax, or wants to…
code-review
Use when user wants code reviewed for correctness, style, bugs, and maintainability. Triggers on: review code, code review, check PR, 審查程式碼, 檢查程式碼. Produces severity-classified findings with a verdict. Do NOT use for security-focused audit (prefer security-audit) or SQL-focused review (prefer sql-review).
plan
Use when user needs a phased implementation plan with requirements, file impact, risks, and alternatives before coding — including clarifying vague or ambiguous requirements first. Triggers on: plan, how should we build, unclear requirements, 規劃, 幫我想方案. Produces a structured plan document. Do NOT use for atomic task…