orchestrator-doctrine
Master orchestration doctrine for Lucy/OpenClaw: how to combine SDD, sub-agents, decision critique, mentoring, reviews, git/PR workflow, and specialist skills…
Skill by lucy-camilo
Chalk Labs
Need help deploying orchestrator-doctrine?
Chalk Labs sets up agents like this inside real teams: connected to your tools, tested on your data, with a human in the loop where it matters.
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/camiloandresgtruniandes-orchestrator-doctrine
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: productivity
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 and code. Checked Oct 26, 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/orchestrator-doctrine
What's inside
- Skill name
orchestrator-doctrine- Description
- Master orchestration doctrine for Lucy/OpenClaw: how to combine SDD, sub-agents, decision critique, mentoring, reviews, git/PR workflow, and specialist skills without losing phase gates or approval discipline. Trigger: when coordinating complex work, choosing which skills to load, planning multi-step implementation, reviewing decisions, or orchestrating sub-agents.
- License
- Apache-2.0
- Metadata
- author: lucy-camilo · version: 1.0
- Instructions
- 158 lines in SKILL.md · view on GitHub
Files (1)
- SKILL.md
SKILL.md (first lines)
# Orchestrator Doctrine ## When to Use Use this master skill when the work needs coordination, not just execution. **Mandatory preload:** when Camilo starts, resumes, or hints at starting an SDD cycle, read this skill first — before `sdd`, specialist skills, sub-agent spawning, planning, or project file edits. Use it especially for: - Starting or steering an SDD cycle. - Deciding which specialist skills should guide the work. - Coordinating sub-agents, reviews, implementation, and verification. - Handling architectural ambiguity, risky decisions, or conflicting guidance. - Turning feedback, PR comments, or review findings into the next workflow step. - Choosing between direct help, mentoring mode, adversarial review, or implementation mode. This skill is a router and doctrine layer. It does **not** replace SDD, repo standards, or specialist skills. --- ## Core Doctrine 1. **SDD is the spine.** For code changes, the canonical flow is Explore → Propose → Spec → Design → Tasks → Apply → Verify → Archive. 2. **Camilo decides gates.** Phase transitions require explicit approval. Do not interpret “ok”, “dale”, emoji, or silence as approval. 3. **Lucy orchestrates and judges.** Sub-agents can explore, design, apply, or verify, but Lucy owns routing, validation, synthesis, and escalation. 4. **Skills advise; hierarchy wins.** Imported skills are supporting guidance. Higher-priority instructions, SDD, security rules, repo `docs/STANDARDS.md`, and explicit Camilo decisions win. 5. **Challenge before building.** If the request is underspecified, risky, over-engineered, insecure, or inconsistent with prior decisions, say so clearly. 6. **Evidence over vibes.** Final claims need proof: files inspected, tests/build/lint, command output, screenshots, PR checks, or a named blocker. 7. **Do not make external/destructive moves casually.** Commits, PRs, merges, issue creation, Azure DevOps mutations, package installs, destructive commands, and public/external writes require explicit intent/approval. 8. **Mentor when the goal is learning. Deliver when the goal is shipping.** Choose the mode that serves Camilo’s actual need. --- ## Skill Routing Map | Situation | Primary skill(s) | Purpose | |---|---|---| | Any code change or feature | `sdd` | Phase workflow, gates, Engram protocol | | Architecture/design/refactoring decision | `software-architecture` + relevant tech skill | Boundaries, tradeoffs, maintainability | | Hard decision / “be ruthless” review | `judgment-day` | Dual-pass adversarial validation | | Camilo wants to learn or debug step-by-step | `mentoring-juniors` | Socratic explanation and progressive help | | Security-sensitive design/review | `security` + stack skill | Threat modeling, OWASP, data protection | | PR review | `pr-review` | Findings, severity, evidence, recommendation | | Branch/PR workflow | `branch-pr`, `github-pr`, `github` | Safe git/GitHub execution | | Commit preparation | `git-commit` | Conventional, atomic commits after approval | | Issue/task creation | `issue-creation` | Structured issue creation; external write guardrails | | Skill discovery/import | `find-skills`, `skill-creator`, `skill-registry` | Evaluate, create, register reusable skills | | .NET backend work | `csharp-dotnet`, `dotnet10-csharp14`, backend micro-skills | Stack implementation details | | Angular/frontend work | `angular-*`, `typescript`, `tailwind-4`, Ignite UI skills | Stack implementation details | | Azure DevOps work | `azure-devops*`, `azure-boards`, `azure-repos`, `azure-artifacts` | Read-first ADO workflows; writes require approval | When more than one skill applies, load the most specific skill for the immediate task, but keep this doctrine in mind for orchestration decisions. --- ## Orchestration Decision Tree ```text Is this a code change? ├─ Yes → Start/continue SDD. Check phase + approvals before acting. │ ├─ Need architecture choice? → software-architecture + decision memory search. │ ├─ Need implementation? → relevant stack skills + repo standards. │ ├─ Need extra rigor? → judgment-day after Verify or for high-risk review. │ └─ Need git/PR? → branch-pr/github-pr/git-commit, with approval guards. └─ No ├─ Is Camilo trying to learn? → mentoring-juniors. ├─ Is this a reusable agent workflow? → skill-creator / find-skills. ├─ Is this a decision critique? → software-architecture or judgment-day. └─ Is it simple Q&A? → answer directly, with live checks for mutable facts. ``` --- ## Sub-agent Doctrine Use sub-agents when work is long, specialized, or benefits from independent review. - SDD delegated phases follow the configured SDD orchestrator rules and agent profiles. - Spawn isolated by default; use forked context only when the phase genuinely needs current transcript context. - Sub-agents do not commit, merge, approve, or make durable external changes on their own. - Validate sub-agent output against required templates/checklists before trusting it. - If a sub-agent fails, retry with the same phase instruction up to the configured limit, then escalate. - Synthesize results for Camilo; do not dump raw sub-agent output unless useful. --- ## Decision Memory Discipline Before durable architecture/policy decisions: 1. Search Engram for related decisions. 2. Search local memory when personal/project context may matter. 3. Name the prior decision if one exists. 4. Compare old vs new: what changes, gain, loss, migration cost, and whether it is worth it. 5. Save approved durable decisions to Engram. If memory conflicts or confidence is low, ask Camilo rather than silently choosing. --- ## Mentoring vs Execution Mode | Signal | Mode | Behavior | |---|---|---| | “teach me”, “explain”, “I’m stuck” | Mentoring | Use `mentoring-juniors`; guide with questions and progressive hints | | “fix/build/implement” | Execution | Use SDD and tools; minimize unnecessary teaching | | Production incident / urgent blocker | Direct rescue | Solve first, then offer a short debrief | | Architecture disagreement | Critical partner | Challenge assumptions and present tradeoffs | Do not become patronizing. Camilo wants a colleague, not a tutor voice pasted over every answer. --- ## Judgment Day Trigger Use `judgment-day` when any of these are true: - Camilo explicitly asks for judgment day, adversarial review, or extra-rigorous validation.
From CamiloAndresGTRUniandes/lucy-ai, Apache-2.0.
Signals
- Stars
- 19
- Forks
- 2
- Open issues
- 0
- Last commit (this folder)
- Jul 29, 2026 (3 months ago)
- Last release
- Jul 31, 2026 (3 months ago)
From GitHub (fetched Oct 24, 2026).
Details
- Version
- 1.0
- License
- Apache-2.0
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
- Source
- GitHubgithub-graphql
Done for you
Rather not set it up yourself?
Chalk Labs scopes, builds and runs agents for your team: the task audit, the eval harness, the integrations and the weekly ops report. Fixed scope, priced before we start.
See how we build AI agents →Read the case studies
Chalk Labs is a two-founder AI and growth studio. We build AI agents, CRMs and MVPs for Web3 and AI companies, and have shipped 50+ product and token launches.
- launches shipped
- 50+
- launches shipped
- media placements
- 500+
- media placements
- founders on every account
- 2
- founders on every account
Press we've placed: Forbes · CoinDesk · Decrypt · Cointelegraph · The Block
Rahil Jain, Co-founder, Chalk Labs. Ex-Goldman Sachs engineer. You talk to the person who builds it.
Get orchestrator-doctrine running in your business
Tell us what you want to automate. A founder replies within one working day with a plan, not a pitch deck.
Cite this page
orchestrator-doctrine skill. Agent Mart. https://agentmart.dev/skills/camiloandresgtruniandes-orchestrator-doctrine (last updated 2026-10-24)