Decision Journal

What it does

When your agent uses it

• Says log this decision
• Were going with
• Pattern from past decisions
• Have I made a call like this before

Example prompts

• “log this decision,”
• “m deciding to,”
• “re going with,”
• “/decision-journal”

Workflow steps

2 steps, taken from the first numbered list in SKILL.md.

1. 1A new decision — the user is committing to something. Capture it now while the reasoning is fresh.
2. 2A review — the user wants to look back at past decisions and pull patterns.

What it can do on your machine

Read from SKILL.md and the folder at commit fdbf39b. It shows what the files ask for, not the result of running them.

• Tool permissions

  Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

  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 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

Decision Journal loads about 1.4k tokens when it runs. Until then it costs about 172 tokens; SKILL.md has 491 words of instructions outside code blocks.

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

~172

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

~1.4k

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

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 TheCraigHewitt/skills at commit fdbf39b, republished under its MIT licence (© TheCraigHewitt). 491 words, ~1,379 tokens.

name

decision-journal

description

Log a decision the user is making (or has made), with full context — the call, the alternatives, the reasoning, the expected outcome, and what would prove it wrong. Then pattern-match against past decisions to flag relevant prior bets. Use this whenever the user says 'log this decision,' 'I'm deciding to,' 'we're going with,' 'pattern from past decisions,' 'have I made a call like this before,' 'decision journal,' or describes a meaningful business or personal call they're committing to. Also use to do a periodic review of past decisions and surface lessons. Output is a markdown decision file in a `decisions/` folder, with cross-links to similar prior decisions.

# [Decision title]

**Date:** YYYY-MM-DD
**Status:** Active / Reversed / Validated / Refuted
**Reversibility:** Easy / Medium / Hard
**Revisit by:** YYYY-MM-DD

## The decision
One sentence.

## Alternatives considered
- [Option A] — why not
- [Option B] — why not
- [Chosen option] — why yes

## What I'm betting on
The key beliefs that make this decision right. Be specific. "The market wants X" is too vague — "AI-native founders in the $1M-$5M range will pay $5K for a 6-week cohort because their existing peer groups don't get AI" is usable.

## What would prove this wrong
- In 3 months: [observable signal]
- In 6 months: [observable signal]
- In 12 months: [observable signal]

## Related past decisions
Links to relevant prior decision files in this folder, with one-line context.

## Notes
Anything else worth capturing — emotional state, who pushed back, what the room felt like.

# Decision review — [window]

## Patterns I see
3–5 patterns across the decisions in this window. Be specific. "You consistently pick the option with more optionality, even when the focused option had higher upside" is useful. "You're a good decision-maker" is not.

## Decisions that worked
For each: what was decided, what bet paid off, what specifically made it right.

## Decisions that didn't
For each: what was decided, what didn't work, what specifically was wrong about the original reasoning. Be honest — don't soften.

## Decisions due for revisit
Any with a "Revisit by" date that has passed and a status still marked Active. List them.

## Lessons worth carrying forward
The 2–3 things that should change about how the user makes decisions in this domain going forward.

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

Files

SKILL.md and 1 other file in cowork/decision-journal of TheCraigHewitt/skills.

• SKILL.md
• evals/evals.json

Open the folder on GitHubat commit fdbf39b