Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →/cs:decide <memo> — Log a decision to two-layer memory via decision-logger. Approved memo becomes durable; raw transcripts kept for reference. Use when the founder has approved a boardroom memo and the decision must become durable company memory — e.g. right after /cs:boardroom concludes.
.claude/skills/alirezarezvani-decide/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | 14% | 0% |
| case-01 | ✗→✓ | ▲ Improved | 217% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 912% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 22% | 0% |
| case-05 | ✗→✓ | ▲ Improved | 10% | 0% |
Command: /cs:decide <memo-path>
Logs the founder's decision via the decision-logger skill. This is the gate where in-session deliberation becomes durable company memory.
/cs:office-hours → /cs:brief → /cs:boardroom → /cs:decide → /cs:execute → /cs:post-mortem
↑ you are hereThe decision-logger skill maintains two layers:
~/.claude/decisions/raw/. Reference only, never feeds back automatically.~/.claude/decisions/approved/. Feeds into future /cs:office-hours and /cs:founder-mode calls.This split prevents the system from "remembering" unresolved debates as if they were decisions.
A board memo file (output of /cs:boardroom).
~/.claude/decisions/approved/<YYYY-MM-DD>-<slug>.md~/company-vault/10-decisions/)markdown# Decision: <title> **Decided:** YYYY-MM-DD **By:** <founder name> **Memo:** <link to boardroom memo> **Brief:** <link to original brief> **Review checkpoint:** YYYY-MM-DD (90d default) ## Decision **Chose:** <option> **Rejected:** <other options + one-line why> ## Success Criteria (binding) - <metric, threshold, timeframe> ## Kill Criteria (binding) - <metric, threshold, action> ## Preserved Dissent - **<dissenter>:** <unresolved concern> - (preserved verbatim; dissent never erased) ## Next Action - `/cs:execute` → 90-day plan due <date> ## Status History - YYYY-MM-DD: APPROVED
The biggest risk in approved decisions is forgetting why someone disagreed. When the kill criteria trigger, the dissent often turns out to have been correct. Preserving it verbatim — not summarized — keeps the company honest at post-mortem time.
/cs:execute <decision> — build the 90-day plan/cs:freeze <decision> <days> — lock if irreversible/cs:post-mortem <decision> — at 90-day checkpointcs-chief-of-staff runs a weekly stale audit:
/cs:post-mortemdecision-loggercs-chief-of-staff../../references/llm-wiki-bridge.mdVersion: 1.0.0
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-02 | fail→pass | 32,192 | 13,249 | -59% | 1 | 1 | 0% | 3,001 | 3,408 | +14% | 0 | 0 | — |
case-20 | pass→pass | 23,435 | 21,960 | -6% | 1 | 1 | 0% | 2,770 | 3,985 | +44% | 0 | 0 | — |
case-01 | fail→pass | 61,794 | 21,492 | -65% | 1 | 1 | 0% | 1,412 | 4,479 | +217% | 0 | 0 | — |
case-03 | fail→pass | 22,664 | 24,842 | +10% | 1 | 1 | 0% | 434 | 4,392 | +912% | 0 | 0 | — |
case-04 | fail→pass | 17,841 | 9,659 | -46% | 1 | 1 | 0% | 2,234 | 2,716 | +22% | 0 | 0 | — |
case-05 | fail→pass | 14,769 | 6,804 | -54% | 1 | 1 | 0% | 1,591 | 1,749 | +10% | 0 | 0 | — |
case-06 | fail→pass | 16,887 | 10,092 | -40% | 1 | 1 | 0% | 1,730 | 1,710 | -1% | 0 | 0 | — |
case-21 | pass→pass | 23,077 | 13,548 | -41% | 1 | 1 | 0% | 2,981 | 3,201 | +7% | 0 | 0 | — |
case-07 | fail→pass | 14,153 | 7,996 | -44% | 1 | 1 | 0% | 1,750 | 1,414 | -19% | 0 | 0 | — |
case-08 | pass→pass | 14,686 | 8,812 | -40% | 1 | 1 | 0% | 1,216 | 1,679 | +38% | 0 | 0 | — |
case-09 | fail→pass | 11,704 | 4,533 | -61% | 1 | 1 | 0% | 1,125 | 1,621 | +44% | 0 | 0 | — |
case-10 | fail→pass | 8,795 | 2,463 | -72% | 1 | 1 | 0% | 554 | 1,329 | +140% | 0 | 0 | — |
case-22 | pass→pass | 22,957 | 15,643 | -32% | 1 | 1 | 0% | 2,676 | 3,084 | +15% | 0 | 0 | — |
case-11 | fail→pass | 13,874 | 7,468 | -46% | 1 | 1 | 0% | 2,174 | 1,294 | -40% | 0 | 0 | — |
case-12 | fail→pass | 14,369 | 9,229 | -36% | 1 | 1 | 0% | 1,612 | 1,706 | +6% | 0 | 0 | — |
case-13 | fail→fail | 8,804 | 7,559 | -14% | 1 | 1 | 0% | 1,290 | 1,331 | +3% | 0 | 0 | — |
case-14 | fail→pass | 15,619 | 2,807 | -82% | 1 | 1 | 0% | 1,749 | 1,368 | -22% | 0 | 0 | — |
case-15 | pass→pass | 17,691 | 5,972 | -66% | 1 | 1 | 0% | 2,068 | 1,826 | -12% | 0 | 0 | — |
case-16 | fail→pass | 25,754 | 2,217 | -91% | 1 | 1 | 0% | 1,731 | 1,226 | -29% | 0 | 0 | — |
case-17 | fail→pass | 11,635 | 1,533 | -87% | 1 | 1 | 0% | 1,108 | 1,112 | +0% | 0 | 0 | — |
case-18 | pass→fail | 7,706 | 2,916 | -62% | 1 | 1 | 0% | 1,197 | 1,408 | +18% | 0 | 0 | — |
case-19 | pass→pass | 19,403 | 14,984 | -23% | 1 | 1 | 0% | 3,224 | 3,553 | +10% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted, and 19 counted toward the lift figure. The other 3 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +59 percentage points is the difference between those two pass rates over the 19 comparable cases. 1 case got worse with the skill loaded, and it is included in that figure.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
| Model | Method | Date | Lift |
|---|---|---|---|
| gemini-3.6-flash | verified | 8/4/2026 | +41% |
Other measured skills in the registry, with their headline benchmark lift.