subagents-orchestration-guide
Guides subagent coordination through implementation workflows. Use when orchestrating multiple agents, managing workflow phases, or determining autonomous execution mode.
65
57%
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 ./skills/subagents-orchestration-guide/SKILL.mdQuality
Discovery
67%Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.
The description has good structural completeness with an explicit 'Use when...' clause, but suffers from abstract, jargon-heavy language that doesn't clearly convey concrete actions. The domain of subagent coordination is somewhat niche, which helps with distinctiveness, but the lack of specific concrete actions (e.g., 'spawns parallel subagents', 'tracks task dependencies', 'merges results') and natural trigger terms weakens its utility for skill selection.
Suggestions
Replace abstract phrases like 'guides subagent coordination' with specific concrete actions such as 'spawns subagents for parallel tasks, tracks task completion, merges results across agents'.
Add more natural trigger terms users might actually say, such as 'parallel tasks', 'delegate subtasks', 'multi-agent', 'break down work', or 'run tasks simultaneously'.
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | Names the domain (subagent coordination, implementation workflows) and some actions (orchestrating, managing workflow phases, determining execution mode), but the actions are fairly abstract rather than concrete. 'Guides subagent coordination' and 'determining autonomous execution mode' are somewhat vague about what specifically happens. | 2 / 3 |
Completeness | Clearly answers both 'what' (guides subagent coordination through implementation workflows) and 'when' (Use when orchestrating multiple agents, managing workflow phases, or determining autonomous execution mode). The explicit 'Use when...' clause with multiple trigger conditions is present. | 3 / 3 |
Trigger Term Quality | Includes some relevant terms like 'subagent', 'agents', 'workflow phases', and 'autonomous execution mode', but these are fairly technical/jargon-heavy. Users are unlikely to naturally say 'orchestrating multiple agents' or 'autonomous execution mode' — more natural terms like 'parallel tasks', 'delegate work', or 'multi-step project' are missing. | 2 / 3 |
Distinctiveness Conflict Risk | The focus on subagent coordination and workflow phases is somewhat distinctive, but terms like 'managing workflow phases' and 'implementation workflows' are broad enough to potentially overlap with project management, task planning, or general workflow skills. The 'subagent' and 'autonomous execution mode' terms help narrow it but the overall scope remains somewhat ambiguous. | 2 / 3 |
Total | 9 / 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 excels at workflow clarity with well-defined multi-step processes, explicit stop points, validation checkpoints, and escalation paths. However, it is excessively verbose—much content is redundant or could be condensed significantly, and the monolithic structure makes it hard to navigate. The actionability is moderate: while call examples and response field specifications are helpful, much of the guidance remains descriptive rather than directly executable.
Suggestions
Reduce verbosity by 40-50%: eliminate redundant explanations (e.g., stop points appear in both the table and the flow descriptions), merge overlapping sections (constraints + principles), and remove information Claude can infer (e.g., 'subagents cannot directly call other subagents—all coordination flows through the orchestrator' is implied by the orchestrator pattern).
Split detailed content into referenced files: move structured response specifications to a `references/response-schemas.md`, information bridging details to `references/bridging-guide.md`, and scale-specific flows to `references/scale-flows.md`, keeping only a summary table in the main skill.
Make call examples more concrete and copy-paste ready by showing actual JSON structures for both the Agent tool invocation and expected response parsing, rather than prose descriptions of what fields to check.
Consolidate the three scale flow descriptions into a single parameterized flow table or decision matrix rather than listing all steps three times with slight variations.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The skill is extremely verbose at ~350+ lines with significant redundancy. Many sections repeat information (e.g., stop points are listed in multiple places, subagent calling patterns are over-explained). The orchestration principles, constraints, and dialogue points sections contain information Claude could infer. The mermaid diagrams add tokens without proportional clarity gains. | 1 / 3 |
Actionability | The skill provides concrete call examples with subagent_type/description/prompt patterns and specific JSON response fields to check. However, much guidance remains procedural description rather than executable commands—the 'call examples' are illustrative templates rather than copy-paste ready, and many sections describe what to do abstractly (e.g., 'information bridging') rather than showing exact implementations. | 2 / 3 |
Workflow Clarity | The multi-step workflows are exceptionally well-sequenced with explicit stop points, validation checkpoints (document-reviewer, code-verifier, design-sync), clear escalation paths, and feedback loops (e.g., integration-test-reviewer needs_revision → return to task-executor). The mermaid diagram for autonomous execution mode clearly shows decision branches and error recovery paths. The 4-step task cycle includes explicit validation before commits. | 3 / 3 |
Progressive Disclosure | There is a single reference to `references/monorepo-flow.md` at the end, but the massive amount of inline content (structured response specs, information bridging details, all three scale flows) could be split into separate reference files. The document is essentially monolithic despite its section headers, with detailed implementation specifics that would benefit from being in linked documents. | 2 / 3 |
Total | 8 / 12 Passed |
Validation
100%Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.
Validation — 11 / 11 Passed
Validation for skill structure
No warnings or errors.
- Repository
- shinpr/claude-code-workflows
- Commit
2e719be
- 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.