AgentStack
Browse Sign in
Browse Why AgentStack Sell Docs
Sign in
SKILL verified MIT Self-run

Deck

skill-control-decks-think-it-through-deck · by control-decks

Supply the shared Think It Through deck rules and mental model whenever the user invokes deck, plays or combines any Think It Through card, invokes help, or asks to use the deck. Stay silent while a card or help utility produces the response; when invoked alone, initialize the deck on the available conversation.

No reviews yet
0 installs
30 views
0.0% view→install

Install

$ agentstack add skill-control-decks-think-it-through-deck

✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.

Security review

✓ Passed

No issues found. Passed automated security review. · v0.1.0 How review works →

  • Prompt-injection patterns
  • Secret / credential exfiltration
  • Dangerous shell & filesystem operations
  • Untrusted network calls
  • Known-malicious package signatures

What it can access

  • Network access No
  • Filesystem access No
  • Shell / process execution No
  • Environment & secrets No
  • Dynamic code execution No

From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.

View the full security report →

Verified badge

Passed review? Show it. Paste this badge into your README, it links to the public security report.

AgentStack Verified badge Links to your public security report.
[![AgentStack Verified](https://agentstack.voostack.com/badges/verified.svg)](https://agentstack.voostack.com/security/report/skill-control-decks-think-it-through-deck)

Reliability & compatibility

Security review passed
0 installs to date
no reviews yet
2mo ago

Declared compatibility

Claude CodeClaude Desktop

Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.

Preview Execution monitoring

We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.

How agent discovery & health will work →
Are you the author of Deck? Claim this listing to set pricing, connect Stripe payouts, and keep 70% of every sale.
Sign up to claim

About

🧩 Think It Through Deck

Treat this as the deck's shared skill, not as a card. Embed the minimum HACP Draft 0.3 resolver so the deck works without the optional HACP plugin.

Deck model

  • Context: Use the relevant conversation and supplied checkpoints.
  • Binding: Resolve what the combo works on: conversation, topic, axis,

current thought, or current result.

  • Map: Maintain Conversation → Topics → Axes silently when useful. Use

short human labels. Reconstruct the map from available context.

  • Working Object: Pass each visible result to the next card, including a

card from another HACP deck.

  • State: Keep the Think mental model local. Share only the Working Object,

Binding, and active HACP controls.

Ordered resolution

Read prose and explicit cards as one ordered stream. Position carries meaning: a leading card acts on following material or the available object; a card after prose consumes that prose; a card between blocks transforms the object at that point.

Before applying an effect:

  1. resolve every played card and its position;
  2. validate inputs, outputs, relations, and active controls;
  3. reject the full stream without effects when it is invalid;
  4. never reorder cards or play a hidden card.

Think cards map to HACP roles:

on-*                    → binding
thinking moves          → operation
to-brief / to-plan      → operation:artifact
explain / with-*        → presentation

Resolve operations in written order. Pass the Working Object across deck boundaries automatically. Binding cards set the Binding at their position. Apply presentation cards to the current final object without changing its substance.

Outcomes and controls

  • Pass success and compatible blocked results to later cards.
  • Preserve pending while approval is missing.
  • Defer cards that need the completed pending result, then resume them after

approval.

  • Apply active controls before tools or domain effects.
  • Let blocking controls win over approval controls.
  • Show persistent controls on every governed response.

The deck never grants authority to mutate, browse, persist, or contact anyone. Those actions remain governed by the user request, higher instructions, tools, and active controls.

further extends the current object by one grounded creative leap. It must not silently become broad ideation, a proposal, or a list of next actions.

Duration

One-shot cards clear after one agent turn. Interview and Grill stay active until complete, stopped, or redirected. A Binding lasts for its combo, including a multi-exchange loop. Outputs last through their confirmation flow.

Never repeat a cleared effect from cadence. Without a played card, respond normally and show no HACP trace.

Display

Show one complete trace when a combo starts. Show compact continuation badges for later Interview or Grill turns. Keep traces and active-control lines outside artifact bodies.

Return the minimum sufficient final result. Do not print intermediate results or explain HACP unless HACP is the subject.

When loaded with a card or help, add no trace or response of your own.

Help

Treat help as a utility. It may explain the deck or recommend normal conversation, one card, or a combo. It never plays a card or recommends domain actions.

Initialization

When invoked alone, respond only:

> 🧩 **THINK IT THROUGH** · Deck initialized · Binding: · Working Object:

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

Install and usage instructions live in the source repository linked above.

Reviews

No reviews yet, be the first.

Versions

  • v0.1.0 Imported from the upstream source.