subagent-driven-development

Use when executing implementation plans with independent tasks in the current session

Quality

—

Does it follow best practices?

Impact

—

No eval scenarios have been run

Securityby

Passed

No known issues

Quality

Content

77%

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

The body is an exceptionally actionable, well-sequenced orchestration skill with concrete scripts, status contracts, and explicit review feedback loops. Its main weaknesses are some repetition that could be tightened and two referenced prompt-template files that are missing from the bundle.

Suggestions

Provide the referenced implementer-prompt.md and task-reviewer-prompt.md bundle files (or remove/guard the links) so the one-level-deep references resolve.

Deduplicate repeated warnings — consolidate the "never HEAD~1" guidance and the pre-judging/"do not flag" guidance into single canonical sections and reference them.

Tighten the longest prose sections (e.g. Constructing Reviewer Prompts) by converting repeated negative guidance into a short checklist.

Dimension	Reasoning	Score
Conciseness	The body is dense with high-value, non-obvious guidance and avoids explaining concepts Claude already knows, but it is long and repeats some emphases (e.g. the "never HEAD~1" warning and pre-judging guidance appear multiple times), so it could be tightened.	2 / 3
Actionability	It gives concrete, executable guidance throughout — exact script invocations ("scripts/review-package BASE HEAD", "scripts/task-brief PLAN_FILE N"), exact git commands, a precise report-file naming convention, and a four-status implementer contract — all copy-paste ready.	3 / 3
Workflow Clarity	The per-task loop is sequenced with explicit decision points, validation checkpoints (spec ✅ + quality approved), re-review feedback loops, ⚠️-item resolution, and a pre-flight plan review, matching the anchor for clear sequences with explicit validation and error-recovery loops.	3 / 3
Progressive Disclosure	It signals references well (a Prompt Templates section and inline links) and offloads real scripts that exist in the bundle, but the referenced implementer-prompt.md and task-reviewer-prompt.md templates are not present in the bundle, so the one-level-deep references are partly broken.	2 / 3
	Total	10 / 12 Passed

Description

58%

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 provides an explicit use-trigger and a defined niche, but its capability statement is abstract and its trigger terms are slightly internal rather than broadly natural. It is clear and usable but not maximally specific or distinctive.

Suggestions

Add one concrete capability phrase so the "what" is more than a single abstract action, e.g. "dispatches one subagent per task, reviews each, and runs a final whole-branch review".

Broaden trigger terms toward natural user phrasing such as "executing a plan", "breaking a plan into tasks", or "running plan tasks in this session".

Sharpen distinctiveness from the sibling skill by contrasting same-session vs. parallel-session execution directly in the description.

Dimension	Reasoning	Score
Specificity	The description names the domain and action ("executing implementation plans with independent tasks") but does not list multiple distinct concrete actions, matching the anchor for naming a domain and some actions without being comprehensive.	2 / 3
Completeness	It has an explicit "Use when..." trigger (the when), but the what ("executing implementation plans") is a single abstract phrase rather than a concrete capability list, leaving it between the explicit-trigger anchor of 3 and the thinner-what anchor of 2.	2.5 / 3
Trigger Term Quality	"executing implementation plans", "independent tasks", and "current session" are relevant but somewhat internal/technical phrasing, lacking the broader natural-term coverage that would justify a 3.	2 / 3
Distinctiveness Conflict Risk	The "independent tasks in the current session" qualifier carves a niche, but it overlaps with the sibling executing-plans skill, so it is somewhat specific yet still at risk of overlapping with a similar skill.	2 / 3
	Total	8.5 / 12 Passed

Validation

93%

Warnings & errors only

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 15 / 16 Passed

Validation for skill structure

Criteria	Description	Result
relative_links	Relative link issues: 2 missing, 1 suspicious	Warning

	Total	15 / 16 Passed

Repository: obra/superpowers
Commit: d884ae0

Reviewed: 4 days ago

Table of Contents

Discovery Implementation Validation

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.