Wins system

The three wins-* skills cover the “hype doc” / “brag document” pattern: capture wins as they happen — honestly, with evidence and calibrated credit — so you can remix them into an artifact when the moment comes. The artifact is only as good as what’s in the log.

How it works

log (in-week) → nudge (Friday) → curate (promo cycle / interview / perf review)

All wins live in a single chronological file at ~/bettersense-work-reflections/wins.md, newest entry first.

---

Skills

wins-log

Triggers: “log a win”, “I just shipped [X]”, “capture this as a win”, “I want to log something”

Structured capture for a win:

Type:         delivery | judgment | leadership | mentorship | recovery | craft | learning | range | culture
Scope:        sprint | quarter | multi-quarter | ongoing
Collaborators: [named, with roles]
Credit:       [honest solo-vs-shared framing]

What happened: [situation — what was at stake, what was hard]
What I did:   [your specific contribution]
Impact:       [quantified results, audience]
Evidence:     [where someone could verify]

The skill pushes back on:

• Vague impact (“improved the dashboard” → for whom, by how much, by when?)
• Inflated solo credit on team work
• Skipping evidence (“even a Slack thread in #design week of April 15 counts”)
• “I can’t quantify it” — usually untrue with one more pass

The Type field is load-bearing. Different audience artifacts need different type mixes. A promo panel cares about delivery + judgment + leadership; interviews want recovery + judgment + range.

---

wins-due

Triggers: “wins nudge”, “what wins did I have this week?”, “anything to log?”

Or via schedule:

/schedule "Every Friday at 4pm, run /bettersense:wins-due and post the list"

Background monitor included

bettersense ships a built-in background monitor (check-wins-cadence) that nudges you automatically — no scheduling setup required. At the start of each Claude Code session it checks whether you’ve logged a win in the last 14 days (configurable via WINS_NUDGE_THRESHOLD_DAYS). If overdue and the cooldown window has passed, Claude gets one notification pointing you to /wins-log. Use wins-due when you want the full category-probe experience; the monitor is for ambient, passive reminders. See Scheduling routines for details.

Weekly cadence check:

• Computes days since last entry
• If ≥ 7 days, nudges with category-specific probes (not generic “anything notable?”)
• Probe examples: “Any judgment calls this week — a useful no, a scope cut, a pushback?”, “Any mentorship — a code review, a career conversation, onboarding?”
• Accepts “honestly, nothing notable this week” without manufacturing a fake win

Why category probes? Memory is selective. “Anything notable?” gets blank stares. “Did you say a useful no?” gets memory hits.

---

wins-curate

Triggers: “I have a promo cycle coming”, “help me prep for interviews”, “perf review self-eval”, “salary negotiation”, “year-in-review”

Reads the wins log and produces an audience-specific artifact. Five modes:

Mode	What it produces	Favored types
promo-packet	Strengths × scope × impact, cited wins per dimension	delivery, judgment, leadership
interview-stories	4-6 STAR-shaped answers with anticipated follow-ups	recovery, judgment, range
perf-review	Balanced self-eval across all type dimensions	breadth across types
salary-case	Business value framing: revenue, retention, cost saved, risk reduced	delivery, impact-quantified
year-in-review	Personal retrospective, trajectory and type-balance read	all types

Operating principles:

• Never invents wins. Curates from the log only. Names gaps (“judgment is thin — strongest available is X, but a panel may want more”) so you know what to build before submission.
• Calibration is non-negotiable. An overclaimed promo packet hurts more than a thinner-but-honest one. Shared-credit wins are labeled as such.
• Resists marketing-speak. Outputs your voice, not “spearheaded transformative initiatives.”

---

Worked example

See Wins curation: a worked example for a full lifecycle walkthrough from first log entry through a promo-packet curate.

---

Things to know

The “log it later” decay is real. Friday nudges help but won’t fully prevent it. Wins captured weeks after the fact are thinner than wins captured in-week.

Artifact quality is bounded by log quality. Six sparse entries produce a sparse promo packet. Twelve rich entries — situation, specific action, quantified impact, real evidence — produce a compelling one.

Composes with perf-review prep. wins-curate (perf-review) reads wins.md plus self-reflect’s growth language plus stakeholder-synthesize’s relationship signals — the three together produce a self-eval with both delivery breadth and relational depth.