orchestrate-sentinel
Manage and execute Sentinel CLI, observability-gateway, test-platform, installer, chart, and docs work through the repository's central .orchestrator run graph.
Skill by caglarkc
Install
Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.
npx agentmarket install skill/caglarkc-orchestrate-sentinel
Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.
Get the weekly diff
The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.
About
Use cases: testing, data, agent-building
Reviews
Write a reviewNo reviews yet. Installed it? Reviews come from verified installers only. How reviews work
Permissions
What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.
- Tools
- No tools found
- Secrets
- None asked for
- Install scripts
- None found
Network hosts it talks to 1
codeload.github.com· found in install command
Files it touches
~/.claude/skills/orchestrate-sentinel
What's inside
- Skill name
orchestrate-sentinel- Description
- Manage and execute Sentinel CLI, observability-gateway, test-platform, installer, chart, and docs work through the repository's central .orchestrator run graph. Use for project planning, PM decomposition, architecture/contracts, implementation, review, verification, integration, resuming prior work, or coordinating Codex/Cursor/Claude agents. Enforces gateway read-only, install-path separation, evidence contracts, risk gates, and Git checkpoints.
- License
- MIT
- Instructions
- 91 lines in SKILL.md · view on GitHub
Agent Skills spec issues
- files: referenced but missing: scripts/auto-push-watch.sh
Files (4)
- SKILL.md
- openai.yaml · agents/
- module-routing.md · references/
- run-patterns.md · references/
SKILL.md (first lines)
# Orchestrate Sentinel Use `.orchestrator` as the source of truth for non-trivial project work. Do not manage multi-step implementation only in conversation memory. ## Start 1. Read repository-root `README.md` and `INSTALL.md`. 2. Read `.orchestrator/PROJECT-STATE.md` and `.orchestrator/SYSTEM.md`. 3. Read `.orchestrator/COMMIT_CONVENTION.md` and the assigned role under `.orchestrator/roles/`. 4. Run `node .orchestrator/bin/orchestrator.mjs discover` when catalog is missing or repository structure changed. 5. Inspect active runs before creating a duplicate run. For module and capability selection, read [module-routing.md](references/module-routing.md). For reusable graph shapes, read [run-patterns.md](references/run-patterns.md). ## Choose workflow ### Answer, audit, or design only Use a read-only `analysis`, `specification`, `architecture-review`, or `pm-planning` item. Do not infer implementation authorization. ### Build or change Create or resume a run. Apply contract-first ordering: ```text PM scope -> architecture/contract -> implement -> review -> verify -> integration -> PM acceptance ``` Add independent review and verification for every item required by `.orchestrator/config.json` risk policy. ### Diagnose Create read-only discovery/reproduction items first. Do not add a fix item unless the request includes fixing or the user later authorizes it. ### Resume Run `validate`, `status`, then inspect `events.jsonl` and `results/`. Reconcile stale `active` ownership before redispatch. ## PM Manager protocol - Preserve the complete approved scope; never collapse Compose, Kubernetes, and COS into one installer story. - Keep the observability gateway read-only unless the user explicitly approves `gateway-write-expansion`. - Treat `agentic/` as reference code, not product source. - Assign role, capabilities, write scopes, risks, approvals, outputs, and exact acceptance to every item. - Keep architecture, implementation, review, verification, and integration ownership separate. - Record durable decisions in `run.decisions`, not only prose. - Create revision items; never rewrite failed history. - Do not mark completion until accepted integration and documentation synchronization exist. ## Dispatch protocol 1. Run `sync` and `status`. 2. Select the first safe batch. 3. Prefer native subagents for bounded independent items when available. 4. Use worktrees for parallel writers when supported; otherwise serialize. 5. Render a platform handoff when native delegation is unavailable: ```bash node .orchestrator/bin/orchestrator.mjs render <run.json> <item-id> --platform cursor ``` 6. Require `.orchestrator/contracts/result.schema.json` output. 7. Record accepted attempt with `record`; do not manually set `done`. 8. After a write item passes its checks, record one atomic Conventional Commit with `type(scope): summary`. `Work-Item` and `Phase` footers are optional. 9. If the commit message is valid, `git push` is allowed. Never force-push and never use `scripts/auto-push-watch.sh`. ## Role routing - `pm-planning`, backlog, scope, acceptance: `roles/pm-manager.md` - architecture, ADR, schema, module boundary: `roles/architecture-manager.md` - code/config/test/docs mutation: `roles/code-implementer.md` - independent diff/contract review: `roles/independent-reviewer.md` - test execution: `roles/verifier.md` - secret/token/approval/gateway-auth: `roles/security-data-reviewer.md` - accepted cross-module assembly: `roles/integration-manager.md` ## Non-negotiable product invariants - Gateway stays read-only; no alert, dashboard, backend write, or Grafana proxy without approval. - CLI talks to Prometheus/Loki/Tempo only through the gateway. - Secrets stay in environment variables, not committed YAML. - Compose, Kubernetes, and COS install paths stay distinct; COS installer is not complete. - `agentic/` is reference, not the shipped product. - Tool approval, timeouts, and redaction are not silently disabled. - Do not claim a check ran when it did not. ## Completion Return the outcome, accepted artifacts, executed checks, commit SHA, remaining risks, and next graph state. Read-only items do not create empty commits. Valid Conventional Commit messages may be pushed with `git push`; never force-push.
From caglarkc/sentinel-coming, MIT.
More from caglarkc/sentinel-coming
- 0.501score
agentic-cli-autocomplete
caglarkc
CLI deneyimini üretim kalitesine taşımak için etkileşim, çıktı, ergonomi ve güvenli varsayılanlar tasarlanırken kullan.
SkillCFreshness: C★ 2Claude Code +9 - 0.501score
agentic-cli-accessibility
caglarkc
CLI deneyimini üretim kalitesine taşımak için etkileşim, çıktı, ergonomi ve güvenli varsayılanlar tasarlanırken kullan.
SkillCFreshness: C★ 2Claude Code +9 - 0.501score
agentic-cli-config-validation
caglarkc
CLI deneyimini üretim kalitesine taşımak için etkileşim, çıktı, ergonomi ve güvenli varsayılanlar tasarlanırken kullan.
SkillCFreshness: C★ 2Claude Code +9 - 0.501score
agentic-cli-alias-command
caglarkc
CLI deneyimini üretim kalitesine taşımak için etkileşim, çıktı, ergonomi ve güvenli varsayılanlar tasarlanırken kullan.
SkillCFreshness: C★ 2Claude Code +9 - 0.448score
babysit-pr
caglarkc
Babysit a GitHub pull request after creation by continuously polling CI checks/workflow runs, new review comments, and mergeability state until the PR is…
SkillDFreshness: D★ 2Claude Code +9 - 0.445score
slides
caglarkc
Build, edit, render, import, and export presentation decks with the preloaded @oai/artifact-tool JavaScript surface through the artifacts tool.
SkillDFreshness: D★ 2Claude Code +9 - 0.445score
spreadsheets
caglarkc
Build, edit, recalculate, import, and export spreadsheet workbooks with the preloaded @oai/artifact-tool JavaScript surface through the artifacts tool.
SkillDFreshness: D★ 2Claude Code +9 - 0.440score
sentinel-phase4-executor
caglarkc
Sentinel CLI Faz 4 — Grafana HTTP bağlantı doğrulama: config, doctor, mock testler, dokümantasyon.
SkillDFreshness: D★ 2Claude Code +9 - 0.440score
sentinel-individual-close-executor
caglarkc
Sentinel CLI bireysel kapanış: REPL’e Grafana doctor özetinin secret-safe enjeksiyonu, config bayrağı, testler ve dokümantasyon.
SkillDFreshness: D★ 2Claude Code +9 - 0.436score
agentic-agent-history-compaction
caglarkc
Uzun oturumlarda konuşma özetleme veya kesme politikası ve korunacak mesajları seçerken kullan.
SkillDFreshness: D★ 2Claude Code +9 - 0.436score
agentic-agent-multi-provider-switch
caglarkc
Oturumlar arası profil ve sağlayıcı değişiminde riskleri ve yeni oturum önerisini tanımlarken kullan.
SkillDFreshness: D★ 2Claude Code +9 - 0.436score
agentic-agent-turn-loop
caglarkc
Ajan tur döngüsü, max_turns, durdurma koşulları ve SIGINT ile iptali tanımlarken kullan.
SkillDFreshness: D★ 2Claude Code +9
Signals
- Stars
- 2
- Forks
- 2
- Open issues
- 0
- Last commit (this folder)
- Aug 15, 2026 (2 months ago)
From GitHub (fetched Oct 24, 2026).
Details
- License
- MIT
- Repository
- github.com/caglarkc/sentinel-coming
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
- Source
- GitHubgithub-graphql
More from caglarkc
- 0.501score
agentic-cli-config-validation
caglarkc
CLI deneyimini üretim kalitesine taşımak için etkileşim, çıktı, ergonomi ve güvenli varsayılanlar tasarlanırken kullan.
SkillCFreshness: C★ 2Claude Code +9 - 0.501score
agentic-cli-accessibility
caglarkc
CLI deneyimini üretim kalitesine taşımak için etkileşim, çıktı, ergonomi ve güvenli varsayılanlar tasarlanırken kullan.
SkillCFreshness: C★ 2Claude Code +9 - 0.501score
agentic-cli-alias-command
caglarkc
CLI deneyimini üretim kalitesine taşımak için etkileşim, çıktı, ergonomi ve güvenli varsayılanlar tasarlanırken kullan.
SkillCFreshness: C★ 2Claude Code +9 - 0.501score
agentic-cli-autocomplete
caglarkc
CLI deneyimini üretim kalitesine taşımak için etkileşim, çıktı, ergonomi ve güvenli varsayılanlar tasarlanırken kullan.
SkillCFreshness: C★ 2Claude Code +9 - 0.461score
coaching-member-boundary
caglarkc
Coach-member akışlarında aktif ilişki doğrulaması, stale snapshot riski, role boundary ve read-only/member-safe ekran davranışlarını koruma skill'i.
SkillDFreshness: D★ 7Claude Code +9 - 0.461score
code-implementation-mode
caglarkc
Bu repoda kod yazan/uygulayan ajan modu; önce kısa etki analizi yapar, sonra rules.md ve ilgili katman kurallarına uyarak doğrudan kodu uygular, açık hata ve…
SkillDFreshness: D★ 7Claude Code +9
Cite this page
orchestrate-sentinel skill. Agent Mart. https://agentmart.dev/skills/caglarkc-orchestrate-sentinel (last updated 2026-10-24)