100 situation-specific prompts and 84 agent-ready skills for engineering managers, directors, and VPs.
- Prompts — copy-paste-and-fill templates for recurring leadership tasks (status updates, roadmaps, reviews, incident comms), each with placeholders, an example, and tuning notes.
- Skills — the same expertise packaged so an agent (Claude Code, Copilot, Cursor, or any LLM) can act: gather inputs, apply structure, and produce the finished artifact. Portable plain-markdown files with a small YAML header.
Ready-made instructions for common leadership tasks. Works with any AI — ChatGPT, Claude, Gemini, Copilot, Cursor. Two ways in:
- Prompt — copy, paste, fill the
[BLANKS]. You drive. - Skill — describe the task; the agent asks what it needs and builds it. The agent drives.
Option 1 — Chat (no setup). Find your task in prompts/ → copy the text under "The Prompt" → paste into your AI → replace the [BLANKS] → send. Read the Tuning Notes at the bottom if you want to adjust.
Option 2 — Coding agent. Clone the repo, point your agent at skills/ (see Installation for your tool), then ask in plain English — "write my exec update from these notes…", "review this PR", "can we take this on next quarter?". It asks for what's missing, then hands you the artifact.
Browse from the terminal:
python tools/skills_cli.py list | search <word> | show <name>
First thing to try: open Exec Status Update, paste it into your AI, fill the blanks. That's the whole loop.
Weekly Comms - 15 prompts
The communication layer of leadership. Status updates, stakeholder emails, launch announcements, incident comms, and more.
| # | Prompt | When to Use |
|---|---|---|
| 01 | Exec Status Update | Monday morning, synthesising team leads' updates for your VP/CTO |
| 02 | Team Weekly Summary | Friday wrap-up for your direct team |
| 03 | Stakeholder Project Update | Keeping PMs, design, and business leads aligned on delivery |
| 04 | Cross-Team Dependency Update | Flagging blockers and handoffs across teams |
| 05 | Internal Launch Announcement | Announcing a shipped feature or service to the company |
| 06 | External Launch Announcement | Customer-facing or public comms for a release |
| 07 | Team Wins Celebration | Highlighting team accomplishments to leadership and peers |
| 08 | Org Change Announcement | Communicating reorgs, new hires, or structural shifts |
| 09 | Incident Stakeholder Comms | Real-time or post-incident updates for non-technical stakeholders |
| 10 | Meeting Recap & Action Items | Turning messy meeting notes into clear follow-ups |
| 11 | Skip-Level Update | Writing upward comms for your boss's boss |
| 12 | Board Engineering Summary | Distilling engineering progress for board-level audiences |
| 13 | Engineering Newsletter | Monthly or bi-weekly update for a non-technical company-wide audience |
| 14 | Vendor Escalation Email | Escalating a blocked vendor issue in writing |
| 15 | Engineering All-Hands Agenda | Planning a quarterly or monthly all-hands meeting |
Planning Artifacts - 14 prompts
The documents that shape what gets built and when. Roadmaps, RFCs, retros, decision records, and operational playbooks.
| # | Prompt | When to Use |
|---|---|---|
| 01 | Quarterly Roadmap | Start of quarter, turning goals into a structured plan |
| 02 | RFC Outline | Proposing a technical change that needs cross-team buy-in |
| 03 | Sprint Retro Summary | Turning retro discussion into actionable themes |
| 04 | Tech Debt Prioritisation | Building a case for paying down debt with a ranked backlog |
| 05 | Build vs Buy Analysis | Evaluating whether to build in-house or use a vendor |
| 06 | Project Pre-Mortem | Identifying risks before a project starts |
| 07 | Capacity Planning | Mapping team bandwidth against committed work |
| 08 | Migration Plan | Structuring a phased approach to a system migration |
| 09 | OKR Drafting | Writing measurable OKRs from vague business goals |
| 10 | Architecture Decision Record | Documenting a technical decision with context and trade-offs |
| 11 | Incident Postmortem | Writing a blameless postmortem from timeline notes |
| 12 | Sprint Planning Breakdown | Breaking epics into well-scoped sprint tickets |
| 13 | Engineering Strategy One-Pager | Communicating what you're optimising for and why, in one page |
| 14 | Team Offsite Agenda | Planning a substantive team offsite that produces real decisions |
People Management - 16 prompts
The human side of leadership. Reviews, career conversations, hiring, and the difficult moments.
| # | Prompt | When to Use |
|---|---|---|
| 01 | Performance Review Draft | Review cycle, synthesising a half or full year of work |
| 02 | One-on-One Prep | Before your weekly 1:1 with a direct report |
| 03 | Feedback Synthesis | Combining peer feedback into a coherent narrative |
| 04 | Career Ladder Draft | Creating or refining engineering level expectations |
| 05 | Job Description | Writing a compelling, specific job posting |
| 06 | Interview Rubric | Creating structured evaluation criteria for interviews |
| 07 | PIP Documentation | Drafting a fair, clear performance improvement plan |
| 08 | Promotion Case | Building a compelling case for a direct report's promotion |
| 09 | Team Health Survey Analysis | Finding patterns and actions from survey results |
| 10 | Onboarding Plan | Creating a structured first 30/60/90 days for a new hire |
| 11 | Skip-Level Meeting Prep | Preparing for 1:1s with your reports' reports |
| 12 | Difficult Conversation Prep | Structuring hard feedback or sensitive discussions |
| 13 | Manager README | Writing a working guide to yourself for your direct reports |
| 14 | Layoff Communication | Individual and team messaging for a workforce reduction |
| 15 | Team Values Workshop | Facilitating a session to define real, usable team values |
| 16 | Staff Engineer Scope Document | Defining what a staff engineer owns and how success is measured |
Incident Management - 12 prompts
The operational backbone of engineering reliability. From alert fires through to organisational learning.
| # | Prompt | When to Use |
|---|---|---|
| 01 | Incident Commander Runbook | You're IC for a live incident and need a structured framework |
| 02 | Real-Time Status Page Update | Writing public status updates during an active incident |
| 03 | Customer Apology Email | Post-incident customer communication that rebuilds trust |
| 04 | War Room Facilitation Guide | Running an effective incident response call |
| 05 | On-Call Handoff | Handing off context at the start/end of an on-call rotation |
| 06 | Runbook Generator | Extracting operational knowledge into step-by-step runbooks |
| 07 | Incident Trend Analysis | Quarterly review of incident patterns and systemic issues |
| 08 | Remediation Tracker | Tracking postmortem action items to completion |
| 09 | Severity Classification Guide | Defining P0-P3 severity levels for your organisation |
| 10 | Game Day Plan | Planning a chaos engineering or incident simulation exercise |
| 11 | Escalation Policy Document | Defining who to call, when, and through what channel |
| 12 | Incident Readiness Review | Auditing preparedness before a launch or high-traffic event |
Architecture - 12 prompts
The technical decisions that shape your systems for years. Design, evaluate, document, and communicate architectural choices.
| # | Prompt | When to Use |
|---|---|---|
| 01 | System Design Document | Before building a new service or system |
| 02 | API Contract Design | Designing a new API (internal or external) |
| 03 | Scalability Assessment | Evaluating whether a system can handle growth |
| 04 | Data Model Design | Designing schemas for a new domain |
| 05 | Caching Strategy | Deciding what to cache, where, and how to invalidate |
| 06 | Observability Strategy | Designing monitoring, logging, and tracing for a system |
| 07 | SLO Definition | Setting Service Level Objectives for your services |
| 08 | Architecture Review Prep | Preparing to present a design for peer review |
| 09 | Dependency Mapping | Documenting service dependencies and blast radius |
| 10 | Technical Vision Document | Writing a long-term technical strategy for your area |
| 11 | Technology Radar | Evaluating and categorising technologies for your org |
| 12 | Database Selection Guide | Choosing the right database for a workload |
Hiring Pipelines - 12 prompts
The end-to-end process of finding, evaluating, and closing engineering candidates.
| # | Prompt | When to Use |
|---|---|---|
| 01 | Hiring Plan | Planning headcount and roles for a quarter |
| 02 | Recruiter Kickoff Brief | Starting a search with a recruiter |
| 03 | Sourcing Outreach Message | Cold outreach to potential candidates |
| 04 | Phone Screen Script | 30-minute initial candidate screen |
| 05 | Take-Home Exercise Design | Creating a fair, well-scoped assessment |
| 06 | Interview Debrief Facilitation | Running a structured hiring decision meeting |
| 07 | Candidate Evaluation Summary | Synthesising interview feedback into a decision |
| 08 | Offer Justification | Building the case for a specific comp package |
| 09 | Candidate Closing Pitch | Selling the role to a finalist who's deliberating |
| 10 | Rejection Email | Delivering a respectful, useful no |
| 11 | Pipeline Analytics Review | Analysing funnel metrics to improve hiring |
| 12 | Interviewer Calibration Guide | Training interviewers for consistency |
Technical Leadership - 12 prompts
For the hands-on leader who still codes — increasingly with AI assistants. Where leadership judgment meets real code: reviewing, orienting, de-risking, estimating, and translating up the chain.
| # | Prompt | When to Use |
|---|---|---|
| 01 | PR Review for Leaders | Reviewing a report's pull request |
| 02 | Codebase Orientation Brief | Getting up to speed on unfamiliar code |
| 03 | Codebase Health Check | A periodic read on a system you own |
| 04 | Prototype Spike Plan | Framing a throwaway experiment before you build it |
| 05 | Spike Readout | Converting a finished spike into a decision |
| 06 | Tech Evaluation Spike | Judging a library/framework/vendor by running it |
| 07 | Legacy Code Assessment | Deciding invest / rewrite / leave on inherited code |
| 08 | Estimation Sanity Check | Before committing a date you didn't produce |
| 09 | Incident Code Triage | Reading a failing code path under pressure |
| 10 | Code to Exec Translation | Explaining a technical change to non-engineers |
| 11 | AI Coding Guidelines | Setting team norms for AI-assisted coding |
| 12 | AI Agent Workflow Design | Wiring AI agents into the dev loop deliberately |
Each file follows the same format:
## Situation - When to use it
## The Prompt - Copy-paste ready, with [PLACEHOLDERS]
## Example Input
## Example Output
## Tuning Notes
Agent-ready versions of the highest-leverage prompts — for when you want the agent to do the task, not just draft text. Each skill lives at skills/<category>/<name>/SKILL.md and follows a portable spec (folded trigger description, Inputs to gather, Steps, Output format, Boundaries, Chaining). See SKILL_TEMPLATE.md to contribute one.
Weekly Comms — 9 skills
| Skill | What it produces |
|---|---|
| exec-status-update | Synthesise messy team-lead updates into a crisp leadership update an exec reads in two minutes |
| team-weekly-summary | A Friday wrap-up for your own engineers: what shipped, what's stuck, what's next |
| stakeholder-project-update | Update non-engineering stakeholders on a project without jargon or false reassurance |
| cross-team-dependency-update | Surface cross-team blockers and handoffs collaboratively, with clear asks and owners |
| meeting-recap-action-items | Turn messy meeting notes into decisions and owned, dated action items |
| board-engineering-summary | Translate a quarter of engineering progress into board-level language and outcomes |
| engineering-newsletter | Turn raw updates into a company-wide newsletter product, design, and sales actually read |
| engineering-all-hands-agenda | A tight all-hands agenda that respects people's time and drives real discussion |
| exec-summarizer | Compress any dense technical material into a headline-first exec summary: bottom line, what matters, what's needed |
Planning — 14 skills
| Skill | What it produces |
|---|---|
| quarterly-roadmap | Goals, backlog, and tech debt into a defensible quarterly roadmap with an explicit not-doing list |
| rfc-outline | A proposed technical change into an RFC thorough enough to decide on |
| sprint-retro-summary | Raw retro feedback distilled into actionable themes with owners |
| tech-debt-prioritisation | A tech-debt list into a prioritised backlog that quantifies the cost of inaction |
| build-vs-buy-analysis | Building in-house vs buying, weighed on total cost and strategic fit |
| project-pre-mortem | Imagine the project failed, then work backward to the likeliest risks and mitigations |
| capacity-planning | Team bandwidth mapped against committed work using realistic, not theoretical, capacity |
| migration-plan | A System A → System B move phased for safety, with rollback at every step |
| okr-drafting | Vague goals into measurable team OKRs with real key results |
| architecture-decision-record | A technical decision captured so future engineers understand the why, not just the what |
| incident-postmortem | A blameless postmortem with systemic root cause and prioritised action items |
| sprint-planning-breakdown | An epic broken into independently deliverable, well-scoped sprint tickets |
| engineering-strategy-one-pager | What engineering is optimising for, and why, on one page |
| team-offsite-agenda | A substantive offsite that produces real decisions, not forced fun |
People Management — 17 skills
| Skill | What it produces |
|---|---|
| performance-review-draft | Scattered observations into a fair, evidence-based review |
| one-on-one-prep | A personalised 1:1 agenda mixing tactical check-ins with career growth |
| feedback-synthesis | Multi-reviewer feedback distilled into a coherent, fair narrative |
| career-ladder-draft | Engineering levels defined with clear, observable expectations |
| job-description | A JD that sells the role honestly and filters effectively |
| interview-rubric | Consistent, bias-resistant evaluation criteria across an interview loop |
| pip-documentation | A fair, evidence-based PIP draft (HR-reviewed) that gives a genuine chance to succeed |
| promotion-case | A case showing the person is already operating at the next level, with evidence |
| compensation-review | A proposed raise positioned against band and compa ratio, with approval flags surfaced |
| onboarding-plan | A 30/60/90 path to a new hire's first meaningful contribution |
| skip-level-meeting-prep | A skip-level that surfaces unfiltered signal and builds trust |
| difficult-conversation-prep | Talking points so hard feedback lands clearly and kindly |
| manager-readme | A head start for reports on your working style and expectations |
| layoff-communication | Coordinated, humane layoff communications for the individual and the team (legal-reviewed) |
| team-values-workshop | A workshop that produces real, usable working principles |
| staff-engineer-scope-doc | What a staff engineer owns, decides, and is measured on |
| team-health-survey-analysis | Survey themes connected to root causes and concrete actions |
Incident Management — 11 skills
| Skill | What it produces |
|---|---|
| incident-commander-runbook | A coordination playbook so anyone on rotation can run an incident calmly |
| severity-classification-guide | An unambiguous P0–P3 guide an engineer can apply at 3am in under a minute |
| war-room-facilitation-guide | A script and structure for the IC to run a live war room |
| on-call-handoff | Exactly what the incoming on-call needs to not be blindsided |
| runbook-generator | Tribal knowledge into a runbook any on-call can follow at 3am |
| incident-trend-analysis | The patterns individual postmortems miss across a quarter |
| remediation-tracker | Scattered postmortem action items tracked to completion with clear ownership |
| incident-stakeholder-comms | A short, plain-language incident update for execs with a committed next-update time |
| game-day-plan | An incident simulation realistic enough to surface real gaps |
| escalation-policy-document | A one-page reference that kills "who do I call next?" ambiguity |
| incident-readiness-review | An audit of whether the team can handle incidents before a high-risk event |
Architecture — 12 skills
| Skill | What it produces |
|---|---|
| system-design-document | A blueprint detailed enough that a senior engineer could build from it |
| api-contract-design | The API contract right before building, with versioning and errors from day one |
| scalability-assessment | Whether the architecture survives 5–10x load, and where the bottlenecks are |
| data-model-design | A schema designed for actual query patterns, not theoretical purity |
| caching-strategy | What to cache, where, how to invalidate, and what happens when the cache is down |
| observability-strategy | The information architecture to go from "it's broken" to "here's the line" fast |
| slo-definition | Reliability targets that are measurable, meaningful, and have an error-budget policy |
| architecture-review-prep | The design — and you — prepared for the tough questions reviewers will ask |
| dependency-mapping | Who depends on what, failure behaviour, and blast radius |
| technical-vision-document | An opinionated 12–24 month vision for where the architecture should head |
| technology-radar | Technologies sorted into Adopt / Trial / Assess / Hold with rationale |
| database-selection-guide | A datastore chosen from the workload, not the marketing |
Hiring Pipelines — 9 skills
| Skill | What it produces |
|---|---|
| hiring-plan | Goals, team, and budget into a defensible hiring plan |
| recruiter-kickoff-brief | The insider context a JD leaves out, handed to a recruiter |
| phone-screen-script | A 30-minute screen that assesses fit and sells the role |
| take-home-exercise-design | A take-home that shows real skill, completable in 2–3 hours, fairly scored |
| interview-debrief-facilitation | A debrief plan that forces an evidence-based, anchoring-resistant decision |
| candidate-evaluation-summary | All interview feedback synthesised into one evidence-based evaluation |
| offer-justification | Interview performance connected to a specific level and comp, defensibly |
| interviewer-calibration-guide | A hands-on session that aligns interviewers and reduces bias |
| pipeline-analytics-review | A diagnosis of where candidates are lost and whether to fix speed or quality |
Technical Leadership — 12 skills
For the hands-on leader who still codes — now more of them, with AI assistants. The intersection of leadership judgment and real code: reviewing a report's work, getting oriented fast, de-risking a call with a spike, and translating code up the chain.
| Skill | What it produces |
|---|---|
| pr-review-for-leaders | A report's PR reviewed for correctness plus scope, risk, and mentoring — comments + a private coaching note |
| codebase-orientation-brief | An unfamiliar repo mapped into what it does, its shape, hot paths, risks, and the questions to ask the team |
| codebase-health-check | A scored, evidence-cited health read on a repo you own, with the highest-leverage fixes |
| prototype-spike-plan | A throwaway spike framed by one question, smallest build, hard timebox, and kill/success criteria |
| spike-readout | A finished spike converted into an answer, evidence, recommendation, and open risks |
| tech-evaluation-spike | A library/framework/vendor evaluated by running it, with an adopt/avoid/need-more verdict |
| legacy-code-assessment | An invest/rewrite/leave call on inherited code, evidence-based, with a reassess trigger |
| estimation-sanity-check | An estimate stress-tested against the code — hidden work, assumptions, a range + confidence |
| incident-code-triage | Ranked hypotheses on a live failure, cheapest check each, and the safest next action |
| code-to-exec-translation | A diff or PR translated into the business framing a non-technical audience acts on |
| ai-coding-guidelines | Tool-neutral team norms for AI-assisted coding: when, review bar, secrets, security, licensing, accountability |
| ai-agent-workflow-design | A staged plan for wiring AI agents into the dev loop: per-stage posture, human gates, metrics |
Skills are plain markdown with a small YAML header, so any LLM agent can read them. Point your agent at the file and it works.
Claude Code
Add to your CLAUDE.md:
skillsDir: skills/
GitHub Copilot
Copy the relevant skill folder into .github/copilot-instructions/.
Cursor
Reference a skill inline via @skills/<name>/SKILL.md, or add the path to your .cursorrules.
Any agent
Point your agent at skills/<category>/<name>/SKILL.md — plain markdown with a small YAML header, readable by any LLM.
Stdlib-only Python helpers in tools/ — no third-party deps, no API keys, runnable under any agent:
| Command | Does |
|---|---|
python tools/skills_cli.py list | search | show | cat |
Browse and emit skills/prompts to stdout |
python tools/gen_manifest.py |
Regenerate manifest.json (machine-readable index) |
python tools/gen_crossmap.py |
Regenerate docs/prompt-skill-map.md |
python tools/gen_integrations.py |
Regenerate docs/INTEGRATIONS.md |
python tools/lint_skills.py |
Lint every SKILL.md against the spec |
python tools/check_repo.py |
Verify links, counts, and generated-doc freshness |
python tools/run_script_tests.py |
Run skill-script golden fixtures |
See docs/INTEGRATIONS.md for per-runtime setup and docs/VARIANTS.md for org-size tuning. CI runs the checks on every push and PR.
Both cover the same engineering-leadership tasks — they differ in how you use them:
prompts/is the companion resource for paste-and-fill workflows: copy the prompt, replace the[PLACEHOLDERS], and run it. Best when you want to drive the output yourself.skills/is for when you want the agent to act — gather the inputs (asking for what's missing instead of guessing), apply the structure, and produce the finished artifact, then offer the natural next step.
Not every prompt became a skill. Thin, transactional one-offs (rejection emails, outreach messages, celebration posts, single announcements) stay as prompts — a template beats an agent there. Substantial, multi-part artifacts became skills.
- V1: 36 prompts across comms, planning, and people management
- V2: Additional categories (incident management, architecture, hiring pipelines)
- V3: Agent-ready skill pack (69 skills, portable across Claude Code / Copilot / Cursor / any LLM)
- V3: Org-size variants convention (
docs/VARIANTS.md) — seeded on capacity-planning, rolling out across skills - V4: CLI to list/search/emit skills from the terminal (
tools/skills_cli.py) - V4: Machine-readable manifest, structure linter, and agnostic CI