This skill should be used when the user says "feature spec teams", "arness code feature spec teams", "team feature spec", "debate this feature", "collaborative feature spec", "spec with agent teams", "multi-agent feature spec", "feature spec debate", or wants to develop a feature idea through structured debate between multiple specialist agents (architects, UX experts, and security specialists) before writing the specification. Uses Claude Code's experimental Agent Teams feature. Requires Agent Teams to be enabled. For standard single-agent feature spec, use arn-code-feature-spec instead.
69
62%
Does it follow best practices?
Impact
—
No eval scenarios have been run
Passed
No known issues
Optimize this skill with Tessl
npx tessl skill review --optimize ./plugins/arn-code/skills/arn-code-feature-spec-teams/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 strong description with excellent trigger term coverage and clear disambiguation from a related skill. Its main weakness is that the concrete capabilities (what the skill actually produces or does beyond 'debate') could be more specific. The explicit listing of trigger phrases and the 'use X instead' guidance are particularly effective for skill selection.
Suggestions
Add more specific output actions, e.g., 'Generates a feature specification document after structured debate rounds covering architecture, UX, and security tradeoffs.'
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | The description mentions 'structured debate between multiple specialist agents (architects, UX experts, and security specialists) before writing the specification' which names the domain and some actions, but doesn't list concrete output actions like 'generates requirements', 'produces spec document', 'identifies tradeoffs'. The core action is somewhat vague ('develop a feature idea'). | 2 / 3 |
Completeness | Clearly answers both 'what' (develops a feature idea through structured debate between specialist agents before writing the specification) and 'when' (explicit trigger phrases and use-case description). Also includes helpful disambiguation ('For standard single-agent feature spec, use arn-code-feature-spec instead'). | 3 / 3 |
Trigger Term Quality | Excellent coverage of trigger terms including many natural variations: 'feature spec teams', 'debate this feature', 'collaborative feature spec', 'spec with agent teams', 'multi-agent feature spec', 'feature spec debate'. These cover both natural user phrases and specific command-like triggers. | 3 / 3 |
Distinctiveness Conflict Risk | Highly distinctive with clear differentiation from the related single-agent skill (arn-code-feature-spec). The multi-agent/teams aspect and specific trigger terms like 'agent teams', 'debate', and 'collaborative' create a clear niche that is unlikely to conflict with other skills. | 3 / 3 |
Total | 11 / 12 Passed |
Implementation
35%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This is a highly detailed orchestration skill for multi-agent feature specification that covers many edge cases and conditional paths. Its main weakness is extreme verbosity — the inline content is far too long for a SKILL.md, with detailed spawn prompt specifications, team composition matrices, and greenfield context handling that should be in reference files. The workflow is logically sound but lacks integrated validation checkpoints, and critical operational details are deferred to external files that aren't provided in the bundle.
Suggestions
Move the team composition matrix, spawn prompt specifications (Step 4), and greenfield context loading details into separate reference files, keeping only a concise summary and file references in SKILL.md — this could reduce the file by 60%+.
Add explicit validation checkpoints within the workflow: verify teammate outputs contain expected sections before proceeding to synthesis, validate the final spec against the template structure before saving.
Remove explanatory content Claude can infer — e.g., the lists of security-relevant terms, UI-related terms, and frontend framework names could be condensed to 'Analyze for security/UI relevance' since Claude knows these domains.
Provide a concrete example of a debate round summary and a synthesized spec excerpt so Claude has a clear model of expected output quality and format.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The skill is extremely verbose at ~300+ lines with extensive detail on team composition matrices, spawn prompt contents, greenfield context loading, animation context, style-brief handling, and numerous conditional branches. Much of this could be condensed or moved to reference files. It explains many things Claude could infer (e.g., how to format user expertise context, what security terms to look for) and includes exhaustive inline detail that bloats the token budget significantly. | 1 / 3 |
Actionability | The skill provides concrete steps and specific file paths to read, but most guidance is procedural description rather than executable code. The few code snippets are limited to JSON config and a bash echo command. The spawn prompt construction, debate facilitation, and spec synthesis are described at a high level without concrete examples of what the actual outputs should look like. It references external files (debate-protocol.md, feature-spec-template.md, greenfield-loading.md) for critical details but these aren't provided in the bundle. | 2 / 3 |
Workflow Clarity | The workflow has clear numbered steps with a logical sequence (check prerequisites → capture idea → classify → spawn team → debate → synthesize). However, validation checkpoints are largely absent — there's no verification that teammates produced valid output, no validation of the synthesized spec against the template, and the debate convergence criteria are deferred entirely to an external file. The error handling section is comprehensive but reads as a flat list rather than integrated checkpoints within the workflow. | 2 / 3 |
Progressive Disclosure | The skill references external files appropriately (debate-protocol.md, feature-spec-template.md, greenfield-loading.md) for detailed protocols, which is good progressive disclosure. However, the SKILL.md itself is a monolithic wall of text with enormous inline detail that should be split into reference files — the team composition matrix, spawn prompt specifications, and greenfield context handling could each be separate references. The bundle files aren't provided, making it impossible to verify the references exist, but the structure suggests too much is kept inline. | 2 / 3 |
Total | 7 / 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 | |
1fe948f
Table of Contents
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.