vpe-advisor
by alirezarezvanivpe-advisor is a VP of Engineering operating-advice skill for startup delivery throughput, hiring funnel health, team structure, and production discipline. Use its references and Python tools to analyze DORA metrics, hiring gaps, manager triggers, on-call practices, and strategic planning tradeoffs.
This skill scores 84/100, which means it is a solid listing candidate for directory users who want an agent to provide VP of Engineering operational guidance with less guesswork than a generic prompt. The repository evidence shows clear triggers, substantive decision frameworks, and runnable scripts, though adoption would be easier with explicit installation and quick-start documentation.
- Strong triggerability: the frontmatter names concrete use cases such as falling sprint velocity, broken engineering hiring, unclear team structure, and tech-lead manager timing.
- Operationally useful: includes four focused references for delivery throughput, hiring funnel, team structure, and production discipline rather than a generic leadership prompt.
- Good agent leverage: three stdlib Python scripts provide deterministic analysis for DORA metrics, hiring funnel gaps, and team-structure recommendations with documented JSON schemas.
- No install command or README is present in the skill path, so users may need to rely on the broader repository conventions for installation.
- The advice is intentionally VPE-scoped and excludes CTO architecture ownership, so it may not fit users looking for deep technical architecture guidance.
Overview of vpe-advisor skill
What vpe-advisor is for
vpe-advisor is a VP of Engineering operating-advice skill for startup engineering leaders, founders, and interim executives who need structured help with delivery throughput, hiring funnel health, team design, and production discipline. It is most useful when the question is not “what architecture should we choose?” but “how should the engineering organization ship, hire, structure teams, and run production reliably?”
Best-fit users and decisions
The vpe-advisor skill fits teams facing practical operating decisions: sprint velocity is dropping, PRs wait too long, incidents are recurring, hiring plans are missing targets, or managers are unsure when to split teams or add engineering management. It is also useful for Strategic Planning because it turns vague organizational concerns into decision areas: DORA metrics, funnel conversion, squad/tribe/chapter structure, manager triggers, on-call sustainability, deployment cadence, and postmortem culture.
What makes it different from a generic prompt
Instead of giving broad leadership advice, vpe-advisor anchors its recommendations in four focused reference files and three deterministic Python tools. The references explain the decision logic; the scripts help calculate delivery throughput, hiring funnel gaps, and team-structure recommendations from JSON inputs. This gives the agent more operational scaffolding than a normal “act as a VP Engineering” prompt.
Important boundaries before installing
vpe-advisor is not a CTO, architecture, product strategy, or HR compliance skill. It can discuss Conway’s Law and operating implications, but architectural ownership belongs elsewhere. It also depends on reasonably accurate inputs: deployment counts, lead time, incidents, funnel stage counts, headcount, manager counts, and work-stream complexity. If you only provide opinions, expect directional advice rather than a useful operating plan.
How to Use vpe-advisor skill
vpe-advisor install and files to inspect first
Install from the repository path with your skill manager, for example:
npx skills add alirezarezvani/claude-skills --skill vpe-advisor
After installation, read SKILL.md first to understand the activation scope and question patterns. Then inspect these files before using the skill in a real planning session:
references/delivery_throughput.mdreferences/engineering_hiring_funnel.mdreferences/eng_team_structure.mdreferences/production_discipline.mdscripts/delivery_throughput_analyzer.pyscripts/eng_hiring_funnel_calculator.pyscripts/eng_team_structure_designer.py
The references explain the management judgment; the scripts show the input schemas and deterministic thresholds.
Inputs that produce better vpe-advisor usage
For strong vpe-advisor usage, bring real operating data rather than a general complaint. Useful inputs include:
- Team size, IC count, EM count, director count, squads, and active work streams
- Deployment frequency, median lead time, MTTR, change failure count, and cycle-time stages
- Hiring target, ATS funnel counts, offer acceptance rate, and median time-to-fill
- Current on-call rotation size, incident frequency, deployment model, and postmortem habits
- Planning horizon, constraints, and what decision must be made now
A weak prompt says: “Our engineering team feels slow.” A stronger prompt says: “We have 22 ICs, 3 EMs, 3 squads, 4 work streams, weekly deploys, 8-day median lead time, PR review wait of 30 hours, and 4 hires needed this quarter. Use vpe-advisor to identify the highest-leverage operating fix.”
Using the included scripts in practice
The scripts are helpful when you can supply structured JSON. Run them locally from the skill folder or copy them into a working directory:
python scripts/delivery_throughput_analyzer.py metrics.json
python scripts/eng_hiring_funnel_calculator.py funnel.json
python scripts/eng_team_structure_designer.py team.json
Each script also runs with an embedded sample if no file is provided, which is useful for learning the schema. Use script output as evidence in the prompt: paste the result into the chat and ask vpe-advisor to convert it into an operating recommendation, risks, and a 30-60-90 day plan.
Prompt pattern for strategic planning
For vpe-advisor for Strategic Planning, frame the request as a decision memo, not a brainstorming session:
“Use vpe-advisor. Context: [company stage, headcount, product pressure]. Data: [delivery, hiring, team structure, production metrics]. Decision needed: [what to change this quarter]. Constraints: [budget, hiring freeze, reliability risk, roadmap deadlines]. Output: diagnose bottlenecks, rank interventions, explain tradeoffs, and propose a 30-60-90 day operating plan.”
This prompt works because it gives the skill enough context to choose between competing VPE levers instead of listing every possible engineering-management practice.
vpe-advisor skill FAQ
Is vpe-advisor suitable for early-stage startups?
Yes, especially once the team is large enough that delivery, hiring, or on-call discipline is no longer handled informally. For a five-person engineering team, the skill can still help diagnose throughput or production habits, but some structure recommendations may intentionally say “do not add process yet.”
When should I not use vpe-advisor?
Do not use vpe-advisor as the primary tool for system architecture, technology selection, compensation policy, legal HR process, or performance management decisions. It can identify operational symptoms, but it should not replace domain experts, employment counsel, or architectural review.
How is this different from asking an AI for VPE advice?
A normal prompt depends heavily on the model’s general knowledge. The vpe-advisor skill gives the agent specific decision frames, healthy ranges, anti-patterns, and script-backed calculations. That matters when you need repeatable diagnosis, not generic leadership language.
Does vpe-advisor require perfect metrics?
No, but it does require honest approximations. If DORA, ATS, or incident data is incomplete, label it clearly. The skill can still reason from partial data, but the best outputs come when estimates are separated from measured facts.
How to Improve vpe-advisor skill
Improve vpe-advisor inputs before asking for advice
The most common failure mode is asking for a recommendation without enough operating data. Before invoking vpe-advisor, assemble a one-page snapshot: current team topology, delivery metrics, hiring funnel, incident/on-call state, and the business constraint driving the decision. This prevents the output from becoming a generic engineering leadership checklist.
Ask for tradeoffs, not just recommendations
VPE work often involves choosing which pain to accept. Ask the skill to compare options such as “add EM,” “split squads,” “reduce WIP,” “tighten incident review,” or “increase sourcing volume.” Strong follow-up prompts include: “What is the hidden cost of this recommendation?” and “What metric should improve first if this is working?”
Iterate with evidence after the first answer
After the first output, feed back what is infeasible, politically hard, or contradicted by data. Example: “We cannot hire an EM this quarter, and the platform team owns three critical services. Revise the plan using only role clarification, WIP limits, and deployment-process changes.” This helps vpe-advisor produce an implementable plan instead of an ideal-state org design.
Extend the skill for your operating model
To improve vpe-advisor locally, add company-specific benchmarks, incident severity definitions, hiring-stage names, team topology constraints, and deployment policies. Keep additions decision-oriented: update references with thresholds, examples, and anti-patterns your organization actually uses. If you add scripts, preserve the current pattern of clear JSON schemas and deterministic output so agents can call them reliably.