auditor

Auditor independiente v4 (Conv 3 — Review+Merge, post pm-redesign 2026-05 Punto 4 + story-closure-gate 2026-05-18). Toma story state=developed (AUTO-HANDOFF /dev-team default; manual opt-in via defer_audit:true) → transition state=developed→reviewing → spawna auditor-{be,fe,agentic} según surface. Phase D NEW: gherkin verification matrix (cada scenario 01-spec.md → test path → status, escribe 06-audit/gherkin-matrix.md). Veredicto: APPROVED | CHANGES_REQUESTED | ESCALATED. Self-fix triviales (lint/typo/format) cap 2 iter. Diseño/security/arch → escala. Cuando todos tickets audit-passed, escribe CHECKPOINTS.md (C1-C5 grid: Code | Spec | Architecture | Cross-cutting | Trace) + AUTO-HANDOFF /pm-{brand} merge. Activa cuando user dice: '/auditor', 'audita story', 'revisa tickets', 'verdict', 'review final', 'CHECKPOINTS'.

Quality

90%

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

Quality

Content

81%Weight 40%Scale 1-5

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

The body excels at actionability and workflow clarity with concrete commands, validation gates, and feedback loops, but loses conciseness to pervasive provenance/version-date clutter and inlines some template/policy content that overlaps referenced SSoT files.

Suggestions

Strip inline provenance citations ('> Origen: process-improvement ...', 'R22 ... origen 2026-05-05', 'v4.1 cement 2026-05-19') or consolidate them into a single changelog section to reduce token noise.

Move the full CHECKPOINTS.md template and self-fix quick-reference table into referenced files under a bundle directory, keeping only a brief pointer inline to avoid duplicating the SSoT in auditor-self-fix-policy.md.

Trim the large Anti-patterns / Anti cross-module pollution lists to the highest-value items and reference the rest, since many entries repeat boundaries already stated in Caso D.

Dimension	Reasoning	Score
Conciseness	The body is action-dense rather than fluff-explaining concepts, but is padded with scattered provenance citations ('> Origen: process-improvement 2026-05-05 R1', 'R22 ... origen 2026-05-05', 'v4.1 cement 2026-05-19') and embedded version dates that could be trimmed.	3 / 5
Actionability	Fully executable guidance throughout: exact bash commands, complete Agent() spawn blocks with verbatim prompts, concrete file paths, and copy-paste yaml snippets covering the common audit cases.	5 / 5
Workflow Clarity	Clear Step 0→5 sequence with explicit validation checkpoints (gate-output.json freshness, 'any_fail=true → BLOCK', artifact-on-disk verification) and feedback loops (Caso B re-audit loop, re-validate after self-fix) for a destructive/batch merge operation.	5 / 5
Progressive Disclosure	Good section structure with a clearly signaled '## Referencias' block pointing to external rule files; minor gaps from inlined content (full CHECKPOINTS.md template, self-fix quick-reference table that duplicates the referenced SSoT, large anti-patterns list) that could live in separate files.	4 / 5
	Total	17 / 20 Passed

Description

100%Weight 40%Scale 1-5

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 is strong: third-person voice, comprehensive concrete actions, explicit trigger phrases, and clear what/when guidance. Its only weakness is dense internal jargon (version codes, dates, process refs) that adds noise but is not penalized by these dimensions.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete actions — 'transition state=developed→reviewing', 'spawna auditor-{be,fe,agentic}', 'gherkin verification matrix', 'escribe CHECKPOINTS.md (C1-C5 grid)' — with comprehensive coverage of the audit lifecycle.	5 / 5
Completeness	Clearly answers both 'what' (audit → verdict → checkpoints → handoff) and 'when' via the explicit 'Activa cuando user dice:' clause with concrete trigger phrases.	5 / 5
Trigger Term Quality	Explicit natural triggers are listed verbatim: '/auditor', 'audita story', 'revisa tickets', 'verdict', 'review final', 'CHECKPOINTS' — covering the slash command and common phrasings a user would say.	5 / 5
Distinctiveness Conflict Risk	Occupies a clear niche — independent story-review/merge auditor in a specific dev process — with distinct triggers and minimal overlap risk with other skills.	5 / 5
	Total	20 / 20 Passed

Validation

81%

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 — 13 / 16 Passed

Validation for skill structure

Criteria	Description	Result
skill_md_line_count	SKILL.md is long (671 lines); consider splitting into references/ and linking	Warning
allowed_tools_field	'allowed-tools' contains unusual tool name(s)	Warning
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	13 / 16 Passed

Repository: alpacapurpura/luana-method
Path: claude/skills/auditor/SKILL.md
Commit: caa08df

Reviewed: ~21 hours 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.