Decide

What it does

When your agent uses it

• Stakes are on the table (X
• Go/no-go) - not for open exploration (/advise
• /think-about-it) and never to relitigate a decision canon has already made (/canon-advise)

Example prompts

• “what should we do”
• “X or Y?”
• “is it worth”
• “/decide”

Requirements

• Pre-approved tools (allowed-tools): Bash, Read, Grep, Glob

Workflow steps

8 steps, taken from the step headings in SKILL.md.

1. 1Decision Guard (route before structuring)
2. 2Frame the Decision
3. 3Parallel Retrieval (one batch)
4. 4Expand the Option Set
5. 5Classify the Decision (this selects the machinery)
6. 6Consequence Table + the Matching Rule
7. 7Reasoning Checks (required - full battery)
8. 8Decision Brief (output)

What it can do on your machine

Read from SKILL.md and the folder at commit b9bea90. It shows what the files ask for, not the result of running them.

• Tool permissions

  Pre-approves these tools, so the agent can use them without asking each time:

  • Bash
  • Read
  • Grep
  • Glob

  From allowed-tools in the SKILL.md frontmatter.

• Runs code

  No scripts in the folder and no shell commands in SKILL.md (its code samples are bash and markdown).

  From the folder's file list and the shell code blocks in SKILL.md.

• Network

  No URLs in SKILL.md.

  From URLs in SKILL.md, links to its own repository left out.

• Credentials

  Names no API keys, tokens, secrets or passwords.

  From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Decide loads about 2.7k tokens when it runs. Until then it costs about 194 tokens; SKILL.md has 939 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it

~194

When it runs · the whole SKILL.md, loaded when a task matches

~2.7k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

• NotePre-approves every shell command (allowed-tools: Bash)SKILL.md

  allowed-tools: Bash, Read, Grep, Glob

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from Abilityai/cornelius at commit b9bea90, republished under its MIT licence (© Abilityai). 939 words, ~2,661 tokens.

name

decide

description

Structure a decision, not just advise on it. Switches modes when the question is "what should we do" - expands the real option set (status quo, defer, pilot, hybrids), classifies the decision type (reversibility, one-shot vs repeated, ruin exposure, risk vs radical uncertainty), applies the MATCHING decision rule (ergodic filter, expected-value decomposition, robustness / minimax regret, value-of-information), and delivers a recommendation with tripwires. KB-grounded in the user's own decision-science frameworks; runs the full reasoning-checks battery. Use when options or stakes are on the table ("X or Y?", "is it worth", go/no-go) - not for open exploration (/advise, /think-about-it) and never to relitigate a decision canon has already made (/canon-advise).

allowed-tools

Bash, Read, Grep, Glob

automation

autonomous

user-invocable

true

argument-hint

<the decision - options, stakes, constraints in natural language>

when_to_use

The question presents a choice between courses of action with real stakes - "should we do X or Y", "go/no-go on Z", "is it worth doing W now". Prefer /advise…

metadata.version

1.0

metadata.created

2026-07-31

metadata.updated

2026-07-31

metadata.author

Ability.ai

metadata.changelog

1.0: Initial version - standalone decision-theoretic sibling of /advise (per the 2026-07-31 self-audit + cross-model consensus: a separate skill, never an…

BRAIN_READ_SCOPE=core                          resources/local-brain-search/run_search.sh "[domain term 1]" --limit 3 --json
BRAIN_READ_SCOPE=core,Books,document-insights  resources/local-brain-search/run_search.sh "[domain term 1]" --limit 5 --json
BRAIN_READ_SCOPE=core                          resources/local-brain-search/run_search.sh "[domain term 2]" --limit 3 --json
BRAIN_READ_SCOPE=core,Books,document-insights  resources/local-brain-search/run_search.sh "[domain term 2]" --limit 5 --json
BRAIN_READ_SCOPE=core                          resources/local-brain-search/run_search.sh "[decision-structure term: ergodicity / optionality / reversibility / regret / explore exploit - match to the decision's shape]" --limit 3 --json
BRAIN_READ_SCOPE=core,Books,document-insights  resources/local-brain-search/run_search.sh "[the same decision-structure term]" --limit 5 --json
resources/local-brain-search/run_connections.sh --hubs --json   # fingerprint - always core; for the attractor check

## Decision: [one-line statement]

**Type:** [reversible? one-shot/repeated? ruin exposure? risk/radical uncertainty] → **Rule applied:** [ergodic filter → EV | robustness/minimax regret | value-of-information]

**Option set considered:** [stated + added options; which were eliminated and by what]

**Consequence table:** [options × driving uncertainties]

**Recommendation:** [the choice, produced by the named rule - and why the losing options lose]

**Tripwires (revisit if):**
- [observable condition → what it changes]
- [date-based checkpoint if the decision was defer/pilot]

**KB grounding:** [[Note 1]], [[Note 2]] - [how each shaped the structure]

[reasoning-checks blocks: Epistemic Inversion · Reference Class · Attractor Check · Provenance Base]

© Abilityai, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

Just SKILL.md in .claude/skills/decide of Abilityai/cornelius.

Open the folder on GitHubat commit b9bea90