richfrem/agent-plugins-skills / plugins/agent-orchestration/.apm/skills/orchestrating-agent-work
orchestrating-agent-work
VerifiedInteractive orchestrator for complex engineering work. Analyzes the task, conducts a diagnostic interview to select the appropriate execution strategy (Directed Acyclic Graph versus Loop, and which Loop type), plans the artifact, and supervises execution and verification. Use for "how should I orchestrate this", "plan this workflow", "should this be a DAG or a loop", "run these jobs in parallel", "delegate this to a worker agent", or "run/resume this graph manifest". Not for single trivial edits or skill self-improvement.
evals/evals.json
View on GitHub[
{
"prompt": "Assess this task and decide whether to run a learning loop, dual-loop delegation, or parallel swarm.",
"should_trigger": true
},
{
"prompt": "Analyze this ambiguous work assignment and route it to the appropriate agent-loop pattern.",
"should_trigger": true
},
{
"prompt": "Evaluate the incoming trigger and manage the shared closure \u2014 seal, persist, retrospective.",
"should_trigger": true
},
{
"prompt": "Launch multiple parallel agents to process these independent tasks across worktrees.",
"should_trigger": true
},
{
"prompt": "Delegate this well-defined task to the inner loop and verify execution with correction loops.",
"should_trigger": true
},
{
"prompt": "Run a red team review on this design document through adversarial review rounds.",
"should_trigger": true
},
{
"prompt": "Which looping or orchestration pattern should I use for a refactoring that has parallel exploration reads and ordered sequential mutations?",
"should_trigger": true
},
{
"prompt": "Help me decide whether to use a dual-loop, a swarm, or a graph for this task.",
"should_trigger": true
},
{
"prompt": "What is the best agent execution strategy for running 50 independent doc conversions in parallel?",
"should_trigger": true
},
{
"prompt": "how should I orchestrate this migration",
"should_trigger": true
},
{
"prompt": "which orchestration pattern fits this refactor",
"should_trigger": true
},
{
"prompt": "plan this workflow across multiple steps",
"should_trigger": true
},
{
"prompt": "should this be a DAG",
"should_trigger": true
},
{
"prompt": "run pytest",
"should_trigger": false
},
{
"prompt": "execute this existing graph manifest",
"should_trigger": true
},
{
"prompt": "Run a batch job across 50 files simultaneously with swarm_run.py.",
"should_trigger": true
},
{
"prompt": "Delegate this task to the inner loop agent right now.",
"should_trigger": true
},
{
"prompt": "Audit this security design with adversarial red-teaming.",
"should_trigger": true
},
{
"prompt": "Delegate the task of updating the user database schema to the inner loop agent.",
"should_trigger": true
},
{
"prompt": "The inner loop finished the Python script \u2014 verify it and loop back if it fails.",
"should_trigger": true
},
{
"prompt": "Use inner/outer agent delegation to implement and verify this feature with correction loops.",
"should_trigger": true
},
{
"prompt": "Run a batch job across 50 files simultaneously with multiple parallel agents.",
"should_trigger": true
},
{
"prompt": "Assess this incoming task and route it to the right loop pattern.",
"should_trigger": true
},
{
"prompt": "Create an ADR documenting the decision to adopt hexagonal architecture.",
"should_trigger": false
},
{
"prompt": "Run a batch formatting job across these 50 markdown files using parallel agents.",
"should_trigger": true
},
{
"prompt": "Partition this checklist into independent sub-tasks and run them concurrently across worktrees.",
"should_trigger": true
},
{
"prompt": "Launch multiple agents simultaneously to process these log files \u2014 they don't depend on each other.",
"should_trigger": true
},
{
"prompt": "Delegate the database schema update to a worker agent and verify the result.",
"should_trigger": true
},
{
"prompt": "Route this ambiguous task to the appropriate loop pattern.",
"should_trigger": true
},
{
"prompt": "Start a new research session and synthesize findings into memory.",
"should_trigger": false
},
{
"prompt": "I have an ambiguous multi-step refactoring workflow that I need to break down into a DAG with token budgets and parallel steps.",
"should_trigger": true
},
{
"prompt": "Help me architect an execution graph with scatter and gather barriers for this code migration pipeline.",
"should_trigger": true
},
{
"prompt": "Compile a graph-manifest.json and Mermaid DAG diagram for this multi-step agent workflow.",
"should_trigger": true
},
{
"prompt": "Design an execution topology with parallel reads and sequential worktree writes for our test suite runner.",
"should_trigger": true
},
{
"prompt": "Architect a read-only multi-source dependency scan DAG that fans out across 4 repos and joins at a sync barrier.",
"should_trigger": true
},
{
"prompt": "Fix a single-line off-by-one bug in calculate_tax() within tax_calculator.py.",
"should_trigger": false
},
{
"prompt": "Execute this existing graph manifest and spawn the worktree sandbox.",
"should_trigger": true
},
{
"prompt": "Run pytest on the test_loop_strategies suite and check for failures.",
"should_trigger": false
},
{
"prompt": "Select between dual-loop and swarm strategy for bulk markdown formatting.",
"should_trigger": true
},
{
"prompt": "What is the git status in the current repository?",
"should_trigger": false
},
{
"prompt": "Execute this existing graph manifest in an isolated worktree and roll back if a node fails.",
"should_trigger": true
},
{
"prompt": "Run the compiled graph-manifest.json with --approval none using the approved manifest sha256.",
"should_trigger": true
},
{
"prompt": "Resume running our task DAG; the parallel read batch and sequential mutation nodes are already planned.",
"should_trigger": true
},
{
"prompt": "Inspect the receipts.jsonl and state.json from the last graph runner execution.",
"should_trigger": true
},
{
"prompt": "Design a new execution graph with scatter and gather barriers for this migration.",
"should_trigger": true
},
{
"prompt": "Run a simple exploratory research search across our documentation.",
"should_trigger": false
},
{
"prompt": "Batch process 100 images in parallel with an agent swarm.",
"should_trigger": true
},
{
"prompt": "Delegate a single bugfix to an inner loop executor.",
"should_trigger": true
},
{
"prompt": "Send this design document to the red team for adversarial review \u2014 iterate until approved.",
"should_trigger": true
},
{
"prompt": "Run a red team review on this architecture, cycling through research, bundle, and critique rounds.",
"should_trigger": true
},
{
"prompt": "Review this security design with a red team agent and continue iterating until it gets an Approved verdict.",
"should_trigger": true
},
{
"prompt": "Assess this incoming task and route it to the appropriate loop \u2014 learning loop, swarm, or dual-loop.",
"should_trigger": true
},
{
"prompt": "The red team approved the design. Delegate implementation to the inner loop agent.",
"should_trigger": true
},
{
"prompt": "Partition the test suite into independent chunks and run them in parallel across worktrees.",
"should_trigger": true
}
]Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 471k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |