Agent skill
delegate
Quick delegation template for sub-agent prompts. Use when assigning work to a sub-agent, before invoking the Agent tool, or when preparing prompts for specialized agents. Provides the WHERE-WHAT-WHY framework. For comprehensive delegation guidance, activate the agent-orchestration how-to-delegate skill.
Install this agent skill to your Project
npx add-skill https://github.com/Jamie-BitFlight/claude_skills/tree/main/.claude/skills/delegate
SKILL.md
Delegation Template
Workflow Reference: See Multi-Agent Orchestration for complete delegation flow with DONE/BLOCKED signaling.
Step 1: Analyze the task. Do you have the "WHERE, WHAT, WHY"?
Step 2: Construct the prompt using the template below.
Template
Your ROLE_TYPE is sub-agent.
[Task Identification - one sentence]
OBSERVATIONS:
- [Factual observations already in your context]
- [Verbatim error messages if applicable]
- [Environment or system state if relevant]
DEFINITION OF SUCCESS:
- [Specific measurable outcome]
- [Acceptance criteria]
- [Verification method]
CONTEXT:
- Location: [Where to look]
- Scope: [Boundaries]
- Constraints: [Hard requirements vs Preferences]
ECOSYSTEM CONTEXT:
- [Session-specific facts the agent cannot find in CLAUDE.md or tool descriptions]
- [Authenticated CLIs, non-obvious doc locations, task-specific access]
YOUR TASK:
1. Run /verify (as completion criteria guide)
2. Perform comprehensive context gathering
3. Form hypothesis → Experiment → Verify
4. Implement solution
5. Only report completion after /verify criteria are met
Authoring guidance (for the orchestrator filling in this template — do not include these annotations in the delivered prompt):
- OBSERVATIONS: Pass-through only — data already in your context (user messages, prior agent reports, command outputs you already received). Include file:line references if already known. Include verbatim error messages, not paraphrased. Do NOT pre-gather data for the agent (e.g., don't run
ruff check .before delegating to a linting agent). Do NOT read, grep, or glob files to find context for the agent — the agent has full tool access and an empty context window; it does its own discovery. No interpretations ("I think"), no assumptions ("probably"). SOURCE: agent-orchestration SKILL.md — Pre-Delegation Verification Checklist section. - DEFINITION OF SUCCESS: The "WHAT". Measurable outcomes the agent can verify. When the agent will produce more than ~1 line of output, instruct it to write results to a file and return only the path — this keeps orchestrator context lean. Example:
Write findings to .claude/reports/NAME-YYYYMMDD.md. Return: STATUS: DONE + file path. - CONTEXT: The "WHERE" and "WHY". Location narrows scope; constraints bound the solution space.
Delegation Rules
Check before sending:
| Rule | Check |
|---|---|
| Formula | Delegation = Observations + Success Criteria + Resources - Assumptions - Micromanagement |
| No HOW | Do NOT tell agent how to implement (e.g., "Change line 42 to X") |
| Constraints OK | DO tell agent constraints (e.g., "Must use the 'requests' library") |
| No Assumptions | Do NOT say "The issue is probably..." |
| Full Scope | If code smell found, instruct agent to audit entire pattern, not single instance |
Quick Checklist
- Starts with
Your ROLE_TYPE is sub-agent. - Contains only factual observations
- No assumptions stated as facts
- Defines WHAT and WHY, not HOW
- Lists resources without prescribing tools
Didn't find tool you were looking for?