원클릭으로 Manus에서 모든 스킬 실행

kirei-sentry

스타0

포크0

업데이트2026년 5월 23일 12:22

Set up production-ready Sentry error + performance monitoring for a project. Detects the framework (Electron, Next.js, Vite/React SPA, Node/Nest, React Native), asks the consent + scope + region decisions, verifies the current SDK API via Ref, then spawns the kirei-sentry agent to produce a framework-specific plan and kirei-forge to implement it — with consent gating, recursive PII scrubbing, CI-only source-map upload, region handling, and the exact CI secrets. Use whenever a user wants to add Sentry, set up error tracking / crash reporting, wire production monitoring, add performance tracing, or integrate Sentry into a new or existing project — even if they don't say "kirei". Invoke with /kirei-sentry; the skill asks the key decisions before working.

설치

Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.

Manus에서 실행

출처

Shironex

Shironex/kirei

GitHub 저장소 열기 Creator 저장소 보기

다운로드

Manus에서 실행

0. PARSE FLAGS

Strip these before proceeding:

Flag	Meaning
`--research-only`	Skip Step 5 (the execute agent). Deliver the plan only.
`--framework <fw>`	Override framework detection. Valid: `electron`, `nextjs`, `vite`, `node`, `nest`, `react-native`.
`--self-hosted`	Target a self-hosted Sentry instead of the cloud SaaS.
`--default-on`	Skip the consent question — use default-ON-with-disclosure (server/SaaS). Otherwise the default recommendation is opt-in/default-OFF.
`--scope <s>`	Skip the scope question. Valid: `errors`, `errors+perf`, `full` (incl. Session Replay).

Any flag the user passes must reach the spawned agent's prompt verbatim.

1. ASK THE DECISIONS

Skip any question already answered by a flag. Otherwise use AskUserQuestion (these change the whole shape of the integration, so ask intentionally):

Consent model — "Should error tracking be opt-in (default OFF) or on by default?"
- Opt-in, default OFF (Recommended for desktop / local-first / privacy-sensitive apps) — nothing is sent until the user enables it; surfaced via a settings toggle (and onboarding step for apps that have one).
- On by default, with disclosure — defensible for server-side / SaaS where there's no end-user device; ship a clear privacy disclosure + easy opt-out.
Scope — "What should it capture?"
- Errors / crashes only — smallest data + privacy surface.
- Errors + performance tracing (Recommended) — adds transactions for bottleneck-hunting; for opt-in apps make perf a separate sub-toggle.
- Errors + performance + Session Replay — maximum visibility, largest privacy + bandwidth footprint.
Hosting / region — "Sentry cloud or self-hosted, and which region?" (US sentry.io / EU de.sentry.io / self-hosted URL). The region matters for source-map upload — getting it wrong 404s in CI.

If the user says "just use sensible defaults", proceed with opt-in/default-OFF · errors+perf (perf as separate opt-in) · cloud, ask region from the DSN later and say so.

2. ANNOUNCE PLAN

One line:

"Running kirei-sentry to design a [framework] Sentry setup ([consent] · [scope] · [hosting/region]) → kirei-forge to implement. Findings to docs/sentry/."

Variants:

--research-only: "… (research only — no implementation)."
Framework still unknown: "… detecting framework first."

3. SPAWN THE RESEARCH AGENT

Spawn the kirei-sentry agent with the Agent tool. It has no session context — include everything.

Prompt structure:

Task: Set up production-ready Sentry for this project. [Plus any extra detail the user gave beyond /kirei-sentry.]

Working directory: [current working directory]

Decisions:
- Consent model: [opt-in default-off | default-on with disclosure]
- Scope: [errors | errors+perf | full incl. replay]
- Hosting / region: [cloud US | cloud EU | self-hosted <url>]

Flags: [--framework <fw> if set] [--self-hosted if set]

Context:
[Framework if known, build tooling, monorepo layout, existing settings/store mechanism, CI release workflow, any DSN already present (region hint), constraints the user mentioned.]

Verify the current SDK API via Ref MCP before prescribing — the API drifts across versions.

Deliver: structured KIREI-SENTRY HANDOFF block + write findings to docs/sentry/ in this repo.

Run in the foreground — you need the plan before implementing.

4. REVIEW FINDINGS

When the agent completes, read its KIREI-SENTRY HANDOFF:

Confirm a findings file was written — Glob docs/sentry/*.md for today. If missing (look for FINDINGS FILE NOT WRITTEN), write it yourself from the handoff via the Write tool at docs/sentry/YYYY-MM-DD-sentry-setup.md.
Spot-check 1-2 paths the plan references actually exist.
Confirm complexity. Electron, multi-process, or any plan touching build config + CI + UI is COMPLEX → kirei-forge. Only a trivial single-entry SPA/Node init is build.

If the agent returned no handoff at all (errored / ran out of budget): tell the user in one sentence, point to anything partial, and offer to retry with a narrower scope. Do not fabricate a plan.

5. SPAWN THE EXECUTE AGENT

Skip if --research-only was passed.

Spawn kirei-forge (or kirei-build only if the handoff is genuinely SIMPLE):

Prompt structure:

Working directory: [current working directory]

Here is the KIREI-SENTRY HANDOFF:
[paste full handoff block]

Findings doc: docs/sentry/[filename]

Implement the setup following the fix order. Critical rules:
- Branch first (never the default branch). Small conventional commits.
- Use the project's package manager to add the Sentry packages into the correct workspace(s).
- The PII scrubber is privacy-critical — write a unit test proving it strips the OS username / home paths from event frames, nested breadcrumb data, and extra/contexts.
- The code must work with the DSN/secrets ABSENT (no init, no crash) — verify that gating.
- Do NOT create the Sentry project, set CI secrets, or touch the Sentry dashboard — those are out of scope (list them for the user).
- Run typecheck + lint + the relevant tests and report results. Do NOT push or open a PR.

Run in the foreground.

6. REPORT TO USER

Summarize:

Framework + decisions, and what the agent designed (1-2 sentences).
What kirei-forge changed (files / branch / commits) — or "research only, no code changes".
The manual checklist the user must still do (pull this from the handoff's out-of-scope section):
1. Create the Sentry project → copy the DSN.
2. Create an org auth token (org:ci scope) and set CI secrets: SENTRY_AUTH_TOKEN, SENTRY_ORG, SENTRY_PROJECT, SENTRY_DSN (+ SENTRY_URL for EU/self-hosted). SENTRY_DSN is required for production telemetry to initialize at all.
3. (Optional) Dashboard + Discord/Slack alert rules in the Sentry UI.
How to verify (the checklist from the plan).
Pointer to docs/sentry/[filename].

RULES

Ask the decisions unless flagged. Consent model especially — opt-in vs default-on is a product/privacy decision, not yours to assume.
Verify the SDK API via Ref. The agent does this; if it skipped it, send it back. Stale init code is the most common Sentry failure.
SENTRY_DSN is the one secret that gates everything. Make sure the final report flags that production builds ship dead without it.
Region is a real footgun. EU/self-hosted source-map upload 404s against the US default — the plan must set the url/SENTRY_URL.
Never push or open PRs from this skill. kirei-forge commits locally; the user pushes. Never touch the user's Sentry dashboard or secrets — those are the user's hands only.
Privacy is non-negotiable. sendDefaultPii: false, a recursive scrubber with a test, no Session Replay unless explicitly requested. If the plan is missing any of these, fix the plan before implementing.

이 저장소의 다른 Skills

같은 저장소

kirei

Shironex/kirei

Orchestrate research + execution for any engineering task. Auto-detects task type (security, ui, refactor, perf, arch, test, migrate, review, debug, data, observability, bundle, license, error, eval, or general) and execute complexity (build vs forge), then spawns the right specialist research agent followed by the right execute agent. Use whenever the user asks to investigate, audit, fix, refactor, debug, optimize, review, or implement anything that benefits from research before code — even if they don't say "kirei". Pass --research-only to skip the execute step. Invoke with /kirei followed by a task description.

2026-05-230

kirei-audit

Shironex/kirei

Audit a codebase for quality and maintainability — code smells, DRY violations, god files, dead code, inconsistent conventions, and unfollowed best practices. Depth-tunable like /kirei-deps; a scout pass sizes how many kirei-refactor agents run in parallel (1 → up to 6) to match the repo. Produces one merged, dependency-ordered cleanup plan, then offers to fix the issues in order via kirei-build / kirei-forge. Use whenever a user asks to audit the codebase, find code smells, hunt technical debt, check for DRY violations, find god files or dead code, surface inconsistencies, run a code-quality / health pass, or plan a refactor — even if they don't say "kirei". Invoke with /kirei-audit; the skill asks which depth before working.

2026-05-210

kirei-chain

Shironex/kirei

Run multiple kirei research agents in parallel against the same target and merge their findings into one combined report. Use when a question needs more than one lens (e.g., "audit this module for security AND performance AND architecture", or "audit observability AND error handling together"). Invoke with /kirei-chain followed by a task description; optionally pass --types security,perf,arch to pin the lens set. Valid lenses: security, ui, refactor, perf, arch, test, data, observability, bundle, error, general.

2026-05-060

kirei-discuss

Shironex/kirei

Conversational pros/cons audit for an idea, feature, or project — before any code is written. Walks the user through problem framing, value, constraints, alternatives, risks, and a recommended next step (kill / spike / build now / wait), then writes a structured decision doc to docs/discuss/. Use whenever the user says "I'm thinking about…", "should we build…", "is X a good idea", "let's talk through…", "what do you think about adding Y", "audit this idea", or wants to think through an approach before committing. Pairs with /kirei — the resulting doc can be passed back via /kirei --findings <path> if the user decides to proceed. Invoke with /kirei-discuss followed by an idea description.

2026-05-060

kirei-templatize

Shironex/kirei

Strip an existing JS/TS repo (Electron, Next.js, Vite, monorepo, etc.) down to a reusable starter template. Analyzes the codebase, splits the strip work into independent phases, and runs them via parallel sub-agents. Asks the user at invocation time about target directory, detection mode (user-marked paths / heuristic / keep-list), execution mode (parallel vs sequential), commit style (small conventional vs one-per-phase), co-author attribution, and whether to seed CLAUDE.md in the new template. Use whenever a user asks to "turn this repo into a template", "strip this for a starter", "make a boilerplate from this project", "templatize", "scaffold from existing repo", or "extract template". Invoke with /kirei-templatize followed by an optional task description.

2026-05-060

kirei-deps

Shironex/kirei

Audit a project's dependencies for safety. Detects the package manager (pnpm/npm/yarn/bun/poetry/uv/cargo/go/bundler), runs its audit, fetches GitHub Dependabot alerts when available, and produces a depth-tunable report — quick CVE counts, safe-bump list, or full ordered upgrade plan with risky-major handoff to /kirei migrate. Use whenever a user asks to audit deps, check for CVEs, find safe upgrades, review Dependabot alerts, run npm/pnpm/yarn/bun audit, check for outdated packages, or plan a dependency hygiene pass — even if they don't say "kirei". Invoke with /kirei-deps; the skill will ask which depth before working.

2026-05-060