Claude Skills by fusebase-dev

Use ONLY when verification passed AND operator explicitly says "prepare deploy" / "draft deploy" / "ship it"; drafts deploy handoff, captures deploy hash + probes + smoke, advises spec DRAFT→DONE flip. Do NOT auto-invoke; operator triggers explicitly.

Use when debugging a deployed app backend. Explains how to fetch build logs and runtime logs using the `fusebase remote-logs` command. Only applicable to apps with a backend/ folder. For local development, use dev-debug-logs skill instead.

Use on first Fusebase Flow install, opening an unfamiliar repo, or after major restructuring; produces durable context map (commands, structure, protected paths, risky boundaries). Do NOT use for routine ticket investigation.

Use when operator starts a new ticket ("ship X", "build Y", "add feature", "fix workflow") and no spec exists; produces spec, clarify questions, acceptance criteria. Do NOT use for code edits, mid-flight reorgs, or after a spec is locked.

ALWAYS load at session start; mandatory, never on-demand. After self-attesting, Read references/<role>.md for your role (Product Owner / AI Developer / Architect / Deploy phase) — don't-list + refusal phrasing. Shared-protocol prohibitions and FR-24 digest stay resident; bodies load from references/shared-protocols.md.

Use when changes touch auth, permissions, secrets, env files, deploy config, external messages, data export/import, production DB writes, or customer-visible behavior; surfaces sensitive-path findings + approval-required list. Do NOT use as general code review.

Use when the operator asks to create, update, import, analyze, or preserve a reusable skill or recurring instruction pattern. Do NOT use for one-off ticket work, ordinary code implementation, problem-catalog incidents, or copying external skill text verbatim.

Use when Product Owner defines smoke prompts or deploy handoffs, and when AI Developer / Deploy phase executes post-deploy smoke. Do NOT use for pre-deploy unit/integration/source gates alone; smoke means proving the operator-visible outcome on the deployed surface with ground-truth diagnostics.

Use when the operator explicitly asks for delegation, subagents, parallel agents, or when an AI Developer has independent T-task slices that can safely run in parallel. Do NOT use for simple edits, immediate blocking work, deploy commands, production side effects, or Product Owner code-writing.

Use when implementing, debugging, or running long tool-using work sessions, or when the operator asks about "token waste", "session cost", "why is this so expensive", or runs "/token-waste-audit" — delivers FR-26's execution-time economy rules (scoped reads, no re-reads of unchanged in-context files, two-strike retry rule, targeted edits) with their quality guards, plus the measurement path. Do NOT use for app-decomposition token economy (product-apps-decomposition owns that), verification po...

Use when code changes need a gate report, operator asks "is this ready?", or a single observed failure needs reproducibility check; runs lint/typecheck/tests, smoke, probes, reproducibility-before-fix. Do NOT approve deploy (that's release-deploy-reporting).

Use before committing a bug fix or improvement, when a fix is non-trivial, or when the same area has been patched before. Operationalizes FR-20 — zoom out to root cause before applying a narrow patch. Do NOT use for trivial typo/format edits, for net-new feature work with no prior code (nothing to zoom out from), or as a substitute for reproduce-before-fix (FR-10 / validation-and-qa).

Use when drafting a spec/ACs for an app feature, planning or building UI views with filters/search/reports, implementing delete/mutation flows, or reviewing feature work — routes to the cross-project behavioral quality patterns (QP-xx) that recurring consumer-app defects proved necessary (URL reflects view state; deletes define cascade/orphan policy; empty/loading/error states; chevron alignment; double-submit guards...). Patterns become spec ACs by ID and carry copy-ready smoke recipes. Do N...

Use ONLY when a business-logic doc exists — docs/<app>/business-logic-index.md (AI-default retrieval index), docs/<app>/business-logic.md (human narrative), or docs/en/business-logic.md (the app-business-docs canonical for Fusebase CLI teams) — before fixes/improvements that touch business behavior. Treats the documented business logic as a guard layer — the fix must not silently break documented behavior. Pairs with FR-20 zoom-out. If none of these docs exist, this skill does nothing (silent...

Use ONLY when docs/audience.md exists (audiences defined during onboarding) OR the operator explicitly asks to optimize a surface for client-facing vs internal use. Steers app surfaces differently — client-facing = simple/guided/trust-first; internal = robust controls/power-features. If docs/audience.md is absent and the operator did not explicitly request the posture check, this skill does nothing (silent no-op) — do NOT create the file. Not for projects without defined audiences.

Use before commit, deploy, or PR merge, or when operator asks "review this diff" / "is this safe?"; reviews diff vs spec, decisions, maintainability, scope, tests, rollback. Do NOT write or fix code — review only.

Use when writing or editing code, adding/changing comments, or before committing a code diff — delivers FR-22's tripwire + retrieval-pointer comment policy at write time. Do NOT use for prose/doc edits, non-code tickets, or as a review gate (that's `code-review`).

ALWAYS apply — Mode A on operator chat output (visual, concrete, brief), Mode B on internal-artifact writes (dense, tabular, front-loaded). Mandatory at session start, never on-demand; rules stay resident, examples lazy-load from references/.

Use when the operator asks for options, variations, alternatives, product/UI directions, or divergent ideation before a spec or decision is locked. Do NOT use for deterministic edits, post-lock implementation changes, simple bug fixes, or AI Developer invention of new product direction.

Use before creating, expanding, or revising any AI-consumed documentation artifact (spec, decisions, tasks, verification-gate, handoff, backlog, problem-catalog, product/business-logic docs, project-internal skills). Classifies whether a persistent doc is needed, which tier applies, and how to minimize context cost. Operationalizes FR-23 — documentation budget. Prevents duplicate rationale, narrative padding, and docs created merely because a template exists. Do NOT use for operator chat, sou...

Use ONLY when the operator runs "/find-wasted-code" (or explicitly asks to "scan the repo for dead-end tool calls / broken links / missing helpers / footgun configs / silent push-through"). Statically scans THIS repo for friction footguns and writes a tracked report to docs/wasted-code/report.md. Manual-trigger ONLY — do NOT auto-invoke; the skill carries disable-model-invocation and no hook wires it. Read-only except the one report write. Findings are review candidates, never auto-fixes. Do ...

Use when the operator runs "/find-wasted-effort" or asks about "wasted effort", "ceremony overhead", "process overhead", "outcome-neutral steps", "are our gate stops worth it", "is the Full lane overkill here", or wants the process-per-outcome sibling of /token-waste-audit. Audits Flow ARTIFACTS ON DISK (gate reports, deploy reports, handoffs, approval artifacts, git log, prevents: annotations) for ceremony that bought no safety outcome — distinct axis from token-waste-audit (tokens-per-rule,...

Use when the operator asks "is Fusebase Flow healthy", "check Fusebase Flow", "did fusebase update break anything", "Fusebase Flow status", "restore Fusebase Flow", or asks whether Fusebase CLI and Fusebase Flow agent files conflict. Runs the read-only health engine, reports layer verdicts (`HEALTHY`, `CLI_LAYER_DRIFT`, `CLI_VERSION_UNSUPPORTED`, `FLOW_LAYER_DRIFT`, `SHARED_MERGE_DRIFT`, `PARTIAL_UPGRADE`, `PUBLISHER_PACKAGING_DRIFT`, `EXCEPTION_IN_EFFECT`, `PARTIAL_UNVERIFIED`, `BROKEN`), an...

Use when something worked before and is now broken (a regression), or the operator asks "when did this break", "find the commit that caused X", "it used to work", "compare to a previous version". Locates where a regression entered by comparing commits / bisecting history. Do NOT use for net-new bugs that never worked, for routine rollback (workflows/git-discipline.md), or as a substitute for reproduce-before-fix (FR-10).

Use when a coding session is getting long, before stopping complex implementation/migration/debugging work, when the operator says "prepare a handoff" / "hand this off" / "continue in a new chat" / "/handoff", or when a long autonomous run needs continuity (Mode run-ledger, announced in chat) — archives the superseded restart state to docs/tmp/handoff/archive/ (paper trail), then writes the active state to docs/tmp/handoff.md so a fresh AI session resumes from the exact current point without ...

Use after spec is drafted to plan implementation; produces decisions, tasks, verification-gate, and implementer handoff. Do NOT write code, do NOT lock decisions on operator's behalf, do NOT run before spec exists.

Use at Specify to classify a ticket Full vs Lightweight after bounded read-only diagnosis, and whenever a change looks ordinary / reversible / low-risk ("small fix", "tweak", "hotfix", "drop pretty-printing", "bump a constant"). Operationalizes FR-21 — ceremony proportional to objective risk. Defines mechanical and semantic assessment, the change-note artifact, one-pass build→verify→deploy, and objective promotion. Sensitive or cross-cutting triggers take the Full lane; an initially unknown c...

Use when launching any long or silent work that could hang — your own probe/script/deploy/fetch-loop/browser-automation, a sub-agent, or a workflow — or when the operator says "the background task hangs", "the agent sits idle", "it never resumes", "is it still running?", or a turn ended waiting on something that never came back. Delivers FR-27's never-launch-bare protocol: bounded/watchdog wrapper, in-turn completion, or BLOCKED-AT return; how to diagnose a suspected hang by activity not 0-by...

Use when planning tasks that touch big source files, when an edit would push a file past the module-size ceiling, when the pre-commit module-size ratchet warns or blocks, or when generating policies/module-size-baseline.txt. Operationalizes FR-25 — module-size ratchet (ceiling default 800; over-ceiling files may shrink, never grow; extraction along a responsibility seam is in-scope, not scope creep). Do NOT use for docs/markdown size (FR-23 owns docs), to force decomposition of an existing mo...

Use when docs/north-star.md exists (the project has been onboarded) OR the operator explicitly asks to check/apply the North Star — steer every task, decision, and fix toward the project's locked vision and flag drift away from it. On the absence path it is a silent no-op by default, with ONE narrow sanctioned output: a single one-time onboarding offer (D5, at most once, only when the operator is clearly doing project work) — so a matcher may load it there for that offer alone. It never creat...

Use when a multi-slice phase has been implemented and the operator wants an independent audit of ALL slices before sign-off, or asks to "audit the phase", "review all slices", "independent check". Spawns a fresh sub-agent that examines every slice against the spec. Do NOT use for a single-diff review (use code-review), for security-only review (use security-permissions-review), or before a phase is implemented.

Use when planning how to structure a product — deciding whether to build one large app or several focused apps. Gives generic decomposition guidance always (reliability + token economy), and steers to the specific app breakdown when docs/<app>/product.md defines one. Do NOT use for single-file edits or for app-internal component structure.

Use when starting a new app and the operator wants product documentation designed before code, OR docs/<app>/product.md exists. Ingests operator research and the North Star, then designs per-app product docs that planning builds from. If no product.md exists and the operator did not ask to create one, this skill does nothing (silent no-op) — do NOT auto-create. Not for mid-implementation edits.

Use when the operator types /product-owner, asks to activate Product Owner, start a Product Owner or PO session, or searches /product in Codex for Fusebase Flow project management. Do NOT use for implementing code; route implementation to the AI Developer role.

Use when the operator wants to set up / optimize THIS project for their vision — onboard the project, capture the North Star, define audience, or run the discovery interview. Triggers on "/onboard", "onboard my project", "set up my project", "capture my vision", "set my north star". Product-Owner-owned. Creates project artifacts (docs/north-star.md, fills AGENTS project-values). Do NOT use for ordinary ticket work, for writing application code, or to auto-run on install — it is operator-trigg...

Use ONLY when verification passed AND operator explicitly says "prepare deploy" / "draft deploy" / "ship it"; drafts deploy handoff, captures deploy hash + probes + smoke, advises spec DRAFT→DONE flip. Do NOT auto-invoke; operator triggers explicitly.

Use on first Fusebase Flow install, opening an unfamiliar repo, or after major restructuring; produces durable context map (commands, structure, protected paths, risky boundaries). Do NOT use for routine ticket investigation.

Use when operator starts a new ticket ("ship X", "build Y", "add feature", "fix workflow") and no spec exists; produces spec, clarify questions, acceptance criteria. Do NOT use for code edits, mid-flight reorgs, or after a spec is locked.

ALWAYS load at session start; mandatory, never on-demand. After self-attesting, Read references/<role>.md for your role (Product Owner / AI Developer / Architect / Deploy phase) — don't-list + refusal phrasing. Shared-protocol prohibitions and FR-24 digest stay resident; bodies load from references/shared-protocols.md.

Use when changes touch auth, permissions, secrets, env files, deploy config, external messages, data export/import, production DB writes, or customer-visible behavior; surfaces sensitive-path findings + approval-required list. Do NOT use as general code review.

Use when the operator asks to create, update, import, analyze, or preserve a reusable skill or recurring instruction pattern. Do NOT use for one-off ticket work, ordinary code implementation, problem-catalog incidents, or copying external skill text verbatim.

Use when Product Owner defines smoke prompts or deploy handoffs, and when AI Developer / Deploy phase executes post-deploy smoke. Do NOT use for pre-deploy unit/integration/source gates alone; smoke means proving the operator-visible outcome on the deployed surface with ground-truth diagnostics.

Use when the operator explicitly asks for delegation, subagents, parallel agents, or when an AI Developer has independent T-task slices that can safely run in parallel. Do NOT use for simple edits, immediate blocking work, deploy commands, production side effects, or Product Owner code-writing.

Use when implementing, debugging, or running long tool-using work sessions, or when the operator asks about "token waste", "session cost", "why is this so expensive", or runs "/token-waste-audit" — delivers FR-26's execution-time economy rules (scoped reads, no re-reads of unchanged in-context files, two-strike retry rule, targeted edits) with their quality guards, plus the measurement path. Do NOT use for app-decomposition token economy (product-apps-decomposition owns that), verification po...

Use when code changes need a gate report, operator asks "is this ready?", or a single observed failure needs reproducibility check; runs lint/typecheck/tests, smoke, probes, reproducibility-before-fix. Do NOT approve deploy (that's release-deploy-reporting).

Use before committing a bug fix or improvement, when a fix is non-trivial, or when the same area has been patched before. Operationalizes FR-20 — zoom out to root cause before applying a narrow patch. Do NOT use for trivial typo/format edits, for net-new feature work with no prior code (nothing to zoom out from), or as a substitute for reproduce-before-fix (FR-10 / validation-and-qa).

Use when the operator asks "is Fusebase Flow healthy", "check Fusebase Flow", "did fusebase update break anything", "Fusebase Flow status", "restore Fusebase Flow", or asks whether Fusebase CLI and Fusebase Flow agent files conflict. Runs the read-only health engine, reports layer verdicts (`HEALTHY`, `CLI_LAYER_DRIFT`, `CLI_VERSION_UNSUPPORTED`, `FLOW_LAYER_DRIFT`, `SHARED_MERGE_DRIFT`, `PARTIAL_UPGRADE`, `PUBLISHER_PACKAGING_DRIFT`, `EXCEPTION_IN_EFFECT`, `PARTIAL_UNVERIFIED`, `BROKEN`), an...