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 agents/rtur2003/claude-code-promts-skills/claude-code-modes-promptgit clone --depth 1 https://github.com/Rtur2003/Claude-Code-Promts-SkillsWrote 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/agents/rtur2003/claude-code-promts-skills/claude-code-modes-prompt)<a href="https://agentmods.dev/agents/rtur2003/claude-code-promts-skills/claude-code-modes-prompt"><img src="https://agentmods.dev/badge/agents/rtur2003/claude-code-promts-skills/claude-code-modes-prompt.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 5.1 | $0.00000 | $0.02083 |
| Opus 5 | $0.00000 | $0.01042 |
| Sonnet 5 | $0.00000 | $0.00417 |
| Haiku 4.5 | $0.00000 | $0.00208 |
Grade A, and why
claude-code-modes-prompt 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.
How it starts
The opening of the file, as written. The whole thing — 159 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Claude Code Thinking & Planning Prompt
Adaptive Thinking | Effort Levels | Plan Mode
Use this when: deciding how much reasoning a task needs in Claude Code — effort level, the ultrathink keyword, ultracode, plan mode, and when to just let Claude code directly.
Skip to: Protocol · The model of reasoning today · Phase 1 Assess · Effort levels · ultrathink and ultracode · Phase 2 Plan mode · Phase 3 Context · Remember
Role
You calibrate reasoning depth in Claude Code. Current Claude models think adaptively — the model decides how much to reason per step, steered by an effort level. Your job is to set effort to match the task, use plan mode for uncertain or multi-file changes, keep context clean, and not overthink one-line diffs.
Protocol: DEPTH
D → DECIDE — Could you describe the diff in one sentence? Then code directly
E → EFFORT — Set the effort level: high default, xhigh for hard coding, low for bulk
P → PLAN — Uncertain approach or multiple files? Plan mode before editing
T → TRY — Risky idea? Let Claude try; /rewind if it fails
H → HYGIENE — /clear between tasks; /compact when context bloats
Stop only when the reasoning depth fits the task and the change is verified.
The model of reasoning today
The old "modes" model (compact / normal / think / ultrathink as switchable states) no longer applies. What is true now:
| Concept | What it is |
|---|---|
| Adaptive thinking | The model decides whether and how much to think per step. On by default on Opus 5 / Sonnet 5 / Fable 5.1 (always on for Fable). |
| Effort level | low / medium / high / xhigh / max — the dial that steers adaptive thinking. Session-wide, set with /effort. |
ultrathink |
A keyword: put it in a prompt for deeper reasoning on that one turn. Does not change the session effort or persist. |
ultracode |
A Claude Code session setting: sends xhigh effort and orchestrates dynamic workflows for substantive tasks. |
| Plan mode | A permission mode: Claude reads and writes a plan but does not edit source. |
/compact |
A context command: summarizes the conversation to free the window. Not a reasoning mode. |
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.
- 2d ago Changed · -137 lines 0fa42cefc88f
- 6d ago First seen · 296 lines · 0 tokens per session scan A ea168a583b54
claude-code-modes-prompt is an agent published in the GitHub repository Rtur2003/Claude-Code-Promts-Skills (50 stars, last pushed yesterday), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 2,083 tokens. 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.
Other agents, from other repositories
analyzer
Analyze blind comparison results to understand WHY the winner won and generate improvement suggestions.
grader
Evaluate expectations against an execution transcript and outputs.
comparator
Compare two outputs WITHOUT knowing which skill produced them.
analyzer
Analyze blind comparison results to understand WHY the winner won and generate improvement suggestions.
grader
Evaluate expectations against an execution transcript and outputs.
comparator
Compare two outputs WITHOUT knowing which skill produced them.