Record growth decision
Skill kimsanguine/signal-to-growth/skills/record-growth-decision
Evidence-driven skills that turn customer signals into measurable growth.
npx -y skills add kimsanguine/signal-to-growth --skill record-growth-decisionAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
2 things to look at
- 12 days oldThe repository was created 12 days ago. New is not bad, but a brand new repository carrying a familiar-sounding name is the shape a typosquat arrives in, and there has been no time for anyone else to find a problem with it.
- 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
What its author says it does
Copied from the file, not written here
Record evidence-backed growth decisions with alternatives, counterevidence, approval, review dates, and append-only outcomes. Use when deciding what to build, hold, test, stop, or review, including 성장 의사결정·decision log·가설 기록.
SKILL.md
2.2 KB, 410 tokens by cl100k_base, as published. Nobody here has run it
Record Growth Decision
Create an auditable decision event that can be reviewed against later outcomes without rewriting history.
Inputs
Require:
- decision context and owner;
- hypothesis;
- valid evidence IDs;
- counterevidence;
- at least one alternative;
- review date;
- success and stop conditions;
- current causal confidence.
Workflow
- Restate the decision in a form a reviewer can accept or reject.
- Link supporting evidence by ID.
- Preserve counterevidence and unresolved unknowns.
- List alternatives, including doing nothing when relevant.
- State expected mechanism, success condition, stop condition, and review date.
- Mark causal confidence as unknown, low, medium, or high.
- Keep status at
draftorawaiting_human_reviewuntil a person decides. - Append the decision event. Never modify an earlier event to make a later result look expected.
- Record later outcomes as new events and link superseded decisions.
Boundaries
- Let the model identify missing evidence, alternatives, and counterarguments.
- Use deterministic checks for IDs, required fields, status transitions, and append-only history.
- Require a person to approve the decision and accept risk.
- Do not equate shipped, launched, or deployed with successful.
- Do not increase confidence because a statement appears in multiple model summaries of the same source.
Outputs
Create or append:
decisions.jsonldecision-summary.mdreview-queue.md
Read references/output-contract.md before writing them.
Stop conditions
Stop when evidence references are invalid, the owner or review date is absent, alternatives are missing, or approval is implied rather than explicit.
Verification
Keep evidence.jsonl and decisions.jsonl together, then run:
python3 scripts/stg.py validate-artifacts artifacts/
Report draft, approved, executed, and outcome-recorded states separately.