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/control-decks/think-it-through/helpnpx skills add control-decks/think-it-through --skill helpgit clone --depth 1 https://github.com/control-decks/think-it-throughWrote 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/control-decks/think-it-through/help)<a href="https://agentmods.dev/skills/control-decks/think-it-through/help"><img src="https://agentmods.dev/badge/skills/control-decks/think-it-through/help.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 | $0.00067 | $0.00648 |
| Opus 5 | $0.00034 | $0.00324 |
| Sonnet 5 | $0.00013 | $0.00130 |
| Haiku 4.5 | $0.00007 | $0.00065 |
Grade A, and why
help 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 3d 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 — 80 lines — stays where its author put it; the contents beside it link to each section on GitHub.
🧩 Think Help
Treat this as the deck's contextual help utility, never as a card.
Resolve the request
- With no argument, inspect the current conversation and recommend zero to three useful interactions.
- With a card name or command, explain its use, default binding, effect, result, duration, limits, and one example.
- With a deck concept, explain it in plain language and tie it to a concrete command.
- With a free-form goal, recommend normal conversation, one card, or a combo.
- Use the current provider's command syntax. Fall back to the portable card name when unknown.
Recommend Continue normally when no card would improve the next agent turn. Never play a card, set a Binding, change the conversation, or recommend domain actions. Leave concrete next actions to next.
Deck guide
Starter hand
distill · discuss · further
grill · recap · propose
Operation cards
distill latest human message → clear thoughts
discuss current thought → developed thought
interview smallest unclear subject → shared understanding
grill current testable idea → verdict and risks
recap available conversation → map and synthesis
propose current open decision → one direction
next latest actionable result → one to three actions
further current Working Object → one grounded creative extension
Binding cards
on-conversation → all available topics and axes
on-topic → one named or current topic
on-axis → one named or current axis
Output cards
to-brief → portable thinking checkpoint
to-plan → execution plan for review
Presentation cards
explain → concise explanation of any HACP result
with-diagrams → smallest useful visual
with-reasoning-map → supported reasoning structure
Think It Through uses Conversation → Topics → Axes only in the cards that
need that model. A binding card selects material for the combo. Adjacent cards
receive the current HACP Working Object, including across decks; the mental
model never crosses that boundary. Operations transform the object and
presentation cards render it without advancing it. A compatible HACP session
adapter must already be active; there is no deck initializer.
What ships with it
2 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 3d ago First seen · 80 lines · 67 tokens per session scan A 4e6c51da0992
help is a skill published in the GitHub repository control-decks/think-it-through (3 stars, last pushed 28d ago), licensed MIT. It adds 67 tokens to every session and 648 once invoked, about $0.0003 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
mindtrain
Configure a private MindTrain instance, query its knowledge catalog, create user-approved training domains and knowledge points from AI dialogue or local reference libraries, and run persistent conversational training through Trainer MCP. Use when the user configures MindTrain, asks what domains or topics exist…
cdc-clear-communication
Draft, rewrite, or audit evidence-based public-health communication using an independently expressed interpretation of the CDC Clear Communication Index. Use for patient explanations, health campaigns, risk and emergency messages, public-safety notices, and educational health material that must make one message and…
distill-textbook
将教学视频与教材按节精炼、跨源整合成私人 Markdown 教材。在用户要求教材蒸馏、 多源教材、视频转笔记、私人教材、lecture notes 合并,或用 faster-whisper 处理教学录像时使用。不适用于依赖推导过程或论证脉络的数学与人文学科。.
coach
Programming mentor with four selectable modes — hardcore (pure Socratic, zero code from the AI), standard (Socratic + pseudo-code/generic examples), architect (AI writes boilerplate, user writes core logic), pair (AI writes the real, complete solution directly, no coaching gate, logs a Learning Log entry only when…
mental-model
Explain a concept, feature, ticket, or implementation in plain language to build the user's mental model — grounded in verified code and sources, with a concrete running example, current-vs-target framing, and an explicit invitation to correct assumptions. Use whenever the user wants to UNDERSTAND how something in a…
technology-learning-judgment
Use when creating, revising, or reviewing technical learning material, a technical proposal, or AI-generated implementation that needs accurate mechanism explanations, honest capability boundaries, and industry alternatives rather than API summaries or inflated claims.