how-to-delegate
Scientific delegation framework for orchestrators — provide observations and success criteria while preserving agent autonomy. Use when assigning work to sub-agents, before invoking the Agent tool, or when preparing delegation prompts for specialist agents.
74
68%
Does it follow best practices?
Impact
Pending
No eval scenarios have been run
Passed
No known issues
Optimize this skill with Tessl
npx tessl skill review --optimize ./plugins/agent-orchestration/skills/how-to-delegate/SKILL.mdQuality
Discovery
89%Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.
This is a well-structured description with a clear 'Use when' clause and distinctive trigger terms targeting a specific niche (agent delegation/orchestration). Its main weakness is that the 'what' portion is somewhat conceptual rather than listing concrete discrete actions — it describes a framework philosophy more than specific operations. Overall it performs well for skill selection purposes.
Suggestions
Add more concrete action verbs to the capabilities portion, e.g., 'Structures delegation prompts, defines success criteria, scopes task boundaries, and formats observation handoffs for sub-agents.'
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | The description names the domain ('scientific delegation framework for orchestrators') and some actions ('provide observations and success criteria', 'preserving agent autonomy'), but it doesn't list multiple concrete, specific actions — it's more of a conceptual framing than a list of discrete capabilities. | 2 / 3 |
Completeness | Clearly answers both 'what' (scientific delegation framework providing observations and success criteria while preserving agent autonomy) and 'when' (explicit 'Use when' clause covering assigning work to sub-agents, before invoking the Agent tool, or preparing delegation prompts). | 3 / 3 |
Trigger Term Quality | Includes strong natural trigger terms that a user or orchestrator would use: 'sub-agents', 'Agent tool', 'delegation prompts', 'specialist agents', 'orchestrators', 'success criteria'. These cover the key vocabulary someone would use when needing this skill. | 3 / 3 |
Distinctiveness Conflict Risk | Highly distinctive niche — focuses specifically on agent delegation patterns and orchestrator-to-sub-agent communication. The triggers ('Agent tool', 'sub-agents', 'delegation prompts') are unlikely to conflict with other skills. | 3 / 3 |
Total | 11 / 12 Passed |
Implementation
47%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This skill provides a thorough, well-structured delegation workflow with clear sequencing and validation checkpoints, but it is severely over-engineered for its purpose. The 10-step process with repeated rationale explanations, mermaid diagrams, and verbose placeholder examples consumes far more tokens than necessary — core delegation principles (pass observations not pre-gathered data, define WHAT not HOW, include success criteria) could be expressed much more concisely. The content would benefit greatly from aggressive trimming while preserving the checklist structure.
Suggestions
Reduce to ~50 lines by removing all 'Why' explanation blocks — Claude understands why success criteria and observations matter. Keep only the actionable checklist items and template.
Remove the mermaid diagrams — the conditional execution and skill-loading flowcharts add visual complexity but little actionable value that a simple bullet list wouldn't provide more efficiently.
Consolidate Steps 8 and 10 into a single pre-delegation checklist to eliminate redundancy between the two verification gates.
Move the detailed Step 9 prompt template and per-step examples into a separate TEMPLATE.md file, keeping SKILL.md as a lean overview with the core principles and checklist structure.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Extremely verbose for what it communicates. The 10-step worksheet with extensive examples, mermaid diagrams, repeated 'Why' explanations, and placeholder templates bloats the content significantly. Much of this (e.g., explaining why pre-gathering is bad multiple times, explaining why success criteria matter) is knowledge Claude already possesses. The core principles could be conveyed in ~30 lines. | 1 / 3 |
Actionability | The skill provides a structured worksheet with concrete templates and examples, but it's fundamentally a meta-process (how to prepare a delegation prompt) rather than executable code. The Agent() invocation syntax and skill loading are concrete, but most steps are fill-in-the-blank templates with example text that varies per situation, making it more of a checklist than executable guidance. | 2 / 3 |
Workflow Clarity | The 10-step sequence is clearly ordered with explicit dependencies between steps. Step 8 provides a pre-flight verification checklist before delegation, and Step 10 includes a final validation gate. The conditional execution flowchart at the top handles the empty-arguments edge case. Feedback loops are present (verify -> fix -> re-verify in Step 8). | 3 / 3 |
Progressive Disclosure | References the agent-orchestration skill at the top for the complete delegation flow, which is good one-level-deep referencing. However, the skill itself is monolithic — all 10 steps with extensive examples are inline rather than being split into a quick-reference overview with detailed steps in a separate file. The mermaid diagrams and repeated examples make the single file very long. | 2 / 3 |
Total | 8 / 12 Passed |
Validation
90%Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.
Validation — 10 / 11 Passed
Validation for skill structure
| Criteria | Description | Result |
|---|---|---|
frontmatter_unknown_keys | Unknown frontmatter key(s) found; consider removing or moving to metadata | Warning |
Total | 10 / 11 Passed | |
- Repository
- Jamie-BitFlight/claude_skills
- Commit
11ec483
- Reviewed
Table of Contents
Is this your skill?
If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.