bmad-dev-auto

One iteration of an unattended development loop. Use when invoked by name.

Quality

56%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

Fix and improve this skill with Tessl

tessl review fix ./src/bmm-skills/4-implementation/bmad-dev-auto/SKILL.md

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 a well-structured, highly actionable orchestration skill with explicit sequencing, concrete commands, and a clear terminal HALT protocol. Its main weaknesses are some over-explained rationale and a progressive-disclosure gap: the overview leans on step files that are not bundled in the reviewed set.

Suggestions

Trim the rationale in the Subagents section (the event-loop explanation) to the essential rule 'invoke subagents synchronously/blocking; never background them — an unattended run has no event loop to resume a yielded turn.'

Ensure the referenced step files (e.g. step-01-clarify-and-route.md) are actually bundled alongside SKILL.md so the one-level-deep disclosure is verifiable and complete.

Consider moving the lengthy HALT folder+id dispatch detail into a dedicated reference file, keeping SKILL.md's HALT section to the essential terminal-state rule and a pointer.

Dimension	Reasoning	Score
Conciseness	The body is mostly lean procedural instruction with concrete commands and templates, but several paragraphs over-explain rationale Claude can infer (e.g. the long Subagents paragraph on event loops and why backgrounded calls stall), matching 'mostly efficient but includes some unnecessary explanation or could be tightened'.	2 / 3
Actionability	It gives fully executable commands ('python3 {project-root}/_bmad/scripts/resolve_customization.py --skill {skill-root} --key workflow'), exact file-resolution rules, and copy-paste skeletal markdown templates, matching the 'fully executable code/commands; copy-paste ready' anchor.	3 / 3
Workflow Clarity	Activation is explicitly sequenced (Steps 1-5), Workflow Execution mandates one-step-at-a-time ordering, and the HALT protocol gives a clear terminal-state path with explicit conditionals (resolved/unresolved/ambiguous, exists/missing) — a clear sequence with explicit checkpoints, and no destructive batch ops are missing validation.	3 / 3
Progressive Disclosure	It follows the right pattern — a concise orchestration/activation overview that delegates detail to per-step files via one clearly signaled reference ('Read fully and follow: ./step-01-clarify-and-route.md') — but that referenced step file is not present in the bundle (references/scripts/assets absent, no step file in the root), so the structure cannot be verified and the disclosure is incomplete.	2 / 3
	Total	10 / 12 Passed

Description

35%

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 correctly uses third person and includes a 'Use when' clause, but its trigger is meta/technical ('invoked by name') rather than natural situation-based keywords, and it lists no concrete actions. It reads as a thin placeholder rather than a genuinely triggerable, specific skill description.

Suggestions

Replace 'Use when invoked by name' with natural situation-based triggers, e.g. 'Use when running an unattended/CI development loop to turn a spec or story into a hardened, reviewable artifact.'

Add 2-3 concrete actions the loop performs (e.g. clarify-and-route, implement, run reviewers) so the 'what' lists multiple specific actions.

Include common phrasings users would actually say ('run dev auto', 'unattended dev loop', 'CI dev iteration') to broaden trigger-term coverage.

Dimension	Reasoning	Score
Specificity	It names a domain ('unattended development loop') and a coarse action ('One iteration'), but lists no concrete sub-actions, matching the 'names domain and some actions, but not comprehensive' anchor rather than the multiple-specific-actions anchor at 3.	2 / 3
Completeness	It states what ('One iteration of an unattended development loop') and includes a 'Use when...' clause, but the when is a weak name-invocation trigger rather than a situation-based explicit trigger, so it does not reach the 'clearly answers both with explicit triggers' anchor at 3.	2 / 3
Trigger Term Quality	The only trigger is 'Use when invoked by name', which is meta/technical guidance rather than natural keywords a user would actually say, matching the 'No natural keywords; technical jargon or overly generic' anchor.	1 / 3
Distinctiveness Conflict Risk	'Unattended development loop' is a recognizable niche and name-only invocation limits auto-trigger conflicts, but the description is thin enough that it could still overlap with other dev-loop skills, matching the 'somewhat specific but could still overlap' anchor.	2 / 3
	Total	7 / 12 Passed

Validation

100%

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 — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository: bmad-code-org/BMAD-METHOD
Path: src/bmm-skills/4-implementation/bmad-dev-auto/SKILL.md
Commit: bb45db4

Reviewed: 1 day 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.