Skills overview
bettersense’s skills are organized into six categories. Skills auto-load when your prompt matches their description, or you can invoke them explicitly with /bettersense:<skill-name>.
The anchor-file pattern
Section titled “The anchor-file pattern”Two files dramatically improve output quality across the whole bundle:
profile.md— who you are. Role, level, company context, communication style, current focus, stack. Created byuser-profile. Read by 8+ skills automatically.strategy/<area>.md— what you’re building. Target problem, approach, SMART metrics, tracks, “not working on” section. Created bystrategy-doc. Read by 5 skills automatically.
Neither is required. Skills work without them — outputs are just more generic and harder to defend. See Profile & strategy for the full setup.
Skills by category
Section titled “Skills by category”Stakeholder management
Section titled “Stakeholder management”| Skill | What it does |
|---|---|
stakeholder-register | Register a new stakeholder (manager, peer, report, team). |
stakeholder-reflect | Guided reflection on a registered stakeholder — question library, cadence-aware, dates entries. |
stakeholder-due | Scan all stakeholder files and surface which question × stakeholder pairs are overdue. |
stakeholder-synthesize | Read accumulated reflections for a stakeholder and surface patterns, contradictions, blind spots — with citations. |
stakeholder-manage | Lifecycle ops: list / edit / re-categorize / archive / rename stakeholders. |
Full stakeholder management docs →
Wins system
Section titled “Wins system”| Skill | What it does |
|---|---|
wins-log | Capture a structured win — type, scope, collaborators, credit framing, situation, action, impact, evidence. Pushes back on vague impact. |
wins-due | Weekly nudge to surface forgotten wins, probing by category (judgment, recovery, mentorship, range). |
wins-curate | Turn the wins log into an artifact: promo packet, STAR interview stories, perf-review self-eval, salary case, year-in-review. |
visibility-sponsorship | The between-formal-moments layer: visibility map, sponsor plan, and an amplification rhythm that gets your logged wins known by the people who decide your career. |
Strategy
Section titled “Strategy”| Skill | What it does |
|---|---|
strategy-doc | Interview-driven strategy doc — target problem, personas, SMART metrics, tracks, “not working on”, counter-metrics. |
impact-audit | Stress-test whether a team’s work is connected to something the business cares about — budget test, death spiral check, goal distance check. Run before strategy-doc or prioritization. |
product-pulse | Dated one-page pulse report — usage, system health, AI signals, counter-metrics, ≤3 follow-up investigations. |
pulse-synthesize | Read accumulated pulse reports over a time window and surface trends, anomalies, regime shifts — with citations. |
user-profile | Anchor file — who you are. Created once, read automatically by 8+ skills. |
tech-strategy-writer | Senior-IC technical strategy doc — diagnosis, direction, first moves, explicit “not doing”. The engineering-owned peer of strategy-doc. |
AI product craft
Section titled “AI product craft”| Skill | What it does |
|---|---|
ai-pm-frameworks | Scoping AI features, build-vs-buy-vs-rule, eval design, model selection, guardrail planning. |
metrics-design | Design the metric tree for an AI feature — north star, leading/lagging, counter-metrics. |
demo-prep | Five-pass prep for high-stakes demos: headline, audience, pre-mortem, script, Q&A. Extra handling for AI-feature demos. |
prioritization-frameworks | Pick the right framework (RICE, ICE, WSJF, MoSCoW, Kano, Cost of Delay) and force hidden assumptions out. |
decision-log | Capture decisions in ADR format with reversibility, rationale, and revisit triggers. Load-bearing for AI decisions tied to model versions. |
self-reflect | Self-reflection on leadership, behavior under pressure, time/energy, fulfillment — chronological private file. |
patterns-watch | Scheduled scan across stakeholder + wins + self files. Surfaces unsolicited patterns with citations. |
model-migration-planner | When a new model ships: blast-radius inventory, eval-before-belief, decision-log revisits, staged rollout with rollback criteria for live AI features. |
People leadership
Section titled “People leadership”| Skill | What it does |
|---|---|
coaching-mode | Ask-not-tell mode for development conversations. GROW frame. Pushes back when you catch yourself prescribing. |
feedback-frameworks | Draft and pressure-test feedback using COIN (Connection → Observation → Impact → Next) + SOLID quality checklist. |
one-on-one-prep | Build 1:1 agendas matched to the relationship type (manager→report, peer, skip-level). |
performance-management | Underperformance ladder — coaching → formalized feedback → PIP — with documentation and conversation scripting. |
hiring-craft | Loop design, rubric-writing with behavioral anchors, structured debrief discipline. |
report-career-architect | 12–18 month growth plan for a report — gaps, sequenced experiences, manager moves, risks. |
report-promo-case | Panel-ready promotion packet with cited evidence, calibrated solo/shared credit, pre-empted objections. |
promo-case-glue | Build promotion packets for glue-heavy engineers; translates glue tasks into technical leadership evidence with manager talking points. |
team-diagnosis | Multi-dimensional team health check across eight dimensions — green/yellow/red with cited evidence and top-3 risks. |
workload-equity | Diagnoses fairness of non-promotable work distribution; surfaces bias signals, single-point dependencies, and rebalancing recommendations. |
engineering-health | Engineering systems diagnostic — SPACE-based productivity read, onboarding time as leading indicator, DevEx friction, platform investment case. |
manage-glue-workers | Coaching & protecting glue-heavy engineers: credit amplification, workload rebalancing, promotion advocacy, skill development. |
psychological-safety | Builds team psychological safety through leader vulnerability, constructive response to upward speech, inclusive facilitation, and rituals like pre-mortems. Diagnoses level and prescribes actions. |
team-charter | Facilitates explicit goals/roles/norms conversation for new/resetting teams. Produces living charter as baseline for recurring check-ins (3X3 Commit step). |
team-check-in | Structured recurring conversation to revisit goals/roles/norms. Surfaces drift and guides course-correction (3X3 Check-In step). |
team-close-gap | Turns alignment gaps into small, targeted behavioral changes with owners and environmental support (3X3 Close step). Closes saying-doing gap through realistic tracked steps. |
team-lifecycle | Guides teams through Tuckman’s stages (forming, storming, norming, performing, adjourning) and resets after reorgs. Phase-specific leader interventions. |
team-style-inventory | Helps teams surface and align on working style preferences (communication, decisions, conflict, meetings). Makes implicit norms explicit, accommodates differences. |
team-workspace | Shared git-backed workspace for collaborative artifacts (charters, strategy, decisions), with a hard wall keeping personal reflections local. Drafts the PR; never publishes for you. |
exec-readout-prep | Prep for decision-seeking exec meetings — the falsifiable ask, room power map from your stakeholder reflections, pre-wiring, objection rehearsal. |
commitments | Ledger of what you’ve promised upward — log, review what’s slipping, renegotiate early, and pressure-test new asks before saying yes. |
Cross-cutting
Section titled “Cross-cutting”| Skill | What it does |
|---|---|
glue-audit | Catalog and quantify non-core (glue) work on yourself or team — time estimates, promotion-value, fairness analysis. |
read-the-room | Retrospective interpretation of a meeting or thread — who held back, performative consensus, subtext. Produces hypotheses, not claims. |
influence-without-authority | Coalition-building, positions vs. interests, sequencing moves. Opinionated about the legitimacy line. |
leadership-os | Managing up/down/across — 1:1s, performance reviews, stakeholder updates, post-mortems, cross-functional conflict. |
premortem | Pre-mortem risk assessment using Critical/Perceived/Undiscussed categories to anticipate failure before it happens. Solo or facilitated with a team. |
weekly | The ~15-minute weekly ritual — wins capture, most-overdue stakeholder reflection, patterns scan, pulse glance in one guided session. The recommended entry point for the whole maintenance loop. |
doctor | Health-check for the memory directory — registry integrity, orphaned files, layout drift, backup status. Read-only by default; every fix needs your OK. |
capture | Log something from the conversation you just had — decision, win, stakeholder observation — drafted from what was said, filed after approval. No re-typing. |
career-retro | Periodic longitudinal retrospective across the whole memory — cited growth deltas vs. your last retro, honest gaps, focus for next period. Persists to self/retros/. |
The agents
Section titled “The agents”Agents are different from skills: they run in a focused context window, useful for tasks that carry a lot of state or span multiple steps.