sales-agent-expert

Senior + arquitecto + CTO del módulo sales_agent post redesign 2026-04. Carga §3 protected surfaces, anti-patterns, decisiones cross-fase, checklist pre-commit. NO carga code vivo (paths/LOC/tests cambian). Use cuando user pida cambio/feature/bug en sales_agent: voz de marca, scheduler/payment tools, observabilidad, callback handler, prompt cache, channel registry, semantic router, closer studio, follow_up engine, eval loop, costo agentes. Triggers: 'modifica sales_agent', 'bug en sales_agent', 'agregar tool al agente', 'agente no cierra', 'suena robótico', 'voz de marca', 'agregar canal sales', 'modificar prompt del specialist', 'closer studio', 'follow-up no dispara', 'el agente repite preguntas', 'PII en trazas', 'costo Kimi', 'DeepSeek alias', 'tier pricing 200k', 'eval goldens'.

Quality

93%

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

88%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.

An exceptionally dense, high-signal domain skill: concrete commands, protected-surface tables, SSoT mappings, and a validated pre-coding workflow with hard stops. Its weaknesses are mild: heavy inline date/commit-sensitive detail, and under-signposted use of the existing reference bundle.

Suggestions

Move the Budget + Outbound Gating detail (plan defaults table, archived PR-2 CONTRACT.md pointer, custom-override mechanics) into a references/ file and keep only the two gate snippets and the S2 wiring summary inline.

Individually name and purpose the remaining reference files (conversation-stages.md, humanization-rules.md, tool-patterns.md) in the Pointers section instead of referring to references/ as an unnamed group, so on-demand loading is discoverable.

Consolidate the dated decision-log entries and commit hashes into a single changelog-style reference (or an 'old patterns' section) to reduce time-sensitive tokens in the always-loaded body.

Dimension	Reasoning	Score
Conciseness	Telegraphic and dense with zero padding or explanation of concepts Claude already knows, but it carries substantial date/commit-hash-sensitive detail inline (plan defaults pricing table, archived PR-2 CONTRACT.md pointer, dated decision log entries) that could be trimmed or moved to a reference file.	4 / 5
Actionability	Fully concrete and executable: copy-paste commands ('cat .claude/rules/anti-duplication.md', 'find $(git rev-parse --show-toplevel)/backend/src -name "<basename>.py"'), exact SSoT paths, runnable Python gate calls with parameters, and a 10-item pre-commit checklist.	5 / 5
Workflow Clarity	'Antes de codear (orden estricto)' is an explicit ordered sequence with validation checkpoints (traza-first gating, RED→GREEN reproducer test, arch tests verde) and STOP/escalate feedback loops, plus a separate numbered anti-duplication procedure with a hard STOP before mirroring.	5 / 5
Progressive Disclosure	references/sales-agent-brand-voice.md is cleanly signaled as 'read on demand' and the bundle files are real and one level deep, but the other three reference files (conversation-stages, humanization-rules, tool-patterns) are only referenced as an unnamed group, and the long Budget + Outbound Gating section is inline detail that could be split out.	4 / 5
	Total	18 / 20 Passed

Description

95%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.

A strong, highly targeted description: it states precisely what it loads (and what it deliberately omits), gives an explicit use-when clause, and enumerates a rich set of natural trigger phrases including user-symptom language. The only soft spot is the role-persona framing ('Senior + arquitecto + CTO'), which adds no concrete capability.

Dimension	Reasoning	Score
Specificity	Lists several concrete capabilities ('Carga §3 protected surfaces, anti-patterns, decisiones cross-fase, checklist pre-commit', 'NO carga code vivo') with minor gaps; the 'Senior + arquitecto + CTO' role framing is not a concrete action, so it falls just below the comprehensive anchor 5.	4 / 5
Completeness	Explicitly answers both what it does (loads protected surfaces, anti-patterns, cross-phase decisions, pre-commit checklist; deliberately does not load live code) and when to use it ('Use cuando user pida cambio/feature/bug en sales_agent' plus a concrete Triggers list).	5 / 5
Trigger Term Quality	Comprehensive natural trigger phrases users would actually say, including symptom language and synonyms: 'modifica sales_agent', 'agente no cierra', 'suena robótico', 'voz de marca', 'follow-up no dispara', 'el agente repite preguntas', 'PII en trazas', 'costo Kimi', 'eval goldens'.	5 / 5
Distinctiveness Conflict Risk	Clear niche scoped to a single module with distinct triggers (closer studio, follow_up engine, channel registry, tier pricing 200k, costo agentes); minimal overlap with any generic skill.	5 / 5
	Total	19 / 20 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: alpacapurpura/luana-method
Path: claude/skills/sales-agent-expert/SKILL.md
Commit: caa08df

Reviewed: 16 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.