How to use the Strategic Alignment Checker agent skill

What is an agent skill, and what does this one do?

An agent skill is a folder of instructions and optional files that an AI agent can load when a task matches the skill's description. This one is licensed under MIT and is built around a single idea: strategy usually fails at the cascade from company goals to team execution, not in the boardroom.

When the skill is active, the agent can run a structured diagnosis. It checks whether the strategy is stated clearly enough to cascade, maps which team goals connect to company goals, looks for orphan goals with no parent, flags conflicting goals, identifies silos where departments optimize locally, and looks for places where leadership's message gets distorted. It can then suggest realignment steps such as cross-functional reviews and quarterly checks.

• Strategy articulation test: ask people from different teams the same question and compare answers.
• Cascade mapping: company OKRs to department, team and individual goals.
• Orphan goal detection: goals with no parent company goal.
• Conflict and coverage checks: goals that work against each other, and company OKRs no team owns.
• Silo and communication gap analysis, plus a realignment protocol and a quarterly alignment score.

Install the strategic-alignment skill

Download the ZIP for the skill from the directory page and unzip it. You have two options depending on where you want the skill available.

For a personal install that works across your projects, unzip the folder into ~/.claude/skills/. For a single project, unzip it into the project's .claude/skills/ folder. The skill folder should contain SKILL.md, scripts/alignment_checker.py and references/alignment-playbook.md.

If you use the Claude apps, you can instead upload the skill under Settings → Capabilities → Skills.

• Personal install: ~/.claude/skills/strategic-alignment/
• Project install: .claude/skills/strategic-alignment/ inside your project
• Claude apps: Settings → Capabilities → Skills, then upload the skill

How do you trigger the skill with natural requests?

You do not need a special command. The skill's description lists the situations it covers, so the agent loads it when your request matches. Plain language works: say that teams are pulling in different directions, that OKRs do not connect, or that you want to look for silos or conflicting OKRs.

Give the agent context about your goals. It works best when you paste or provide your company and team OKRs, or point it to a JSON file of them.

• "Our teams seem to be pulling in different directions. Can you help me find where the strategy breaks down?"
• "Check whether each of our team OKRs connects to a company OKR and flag any orphan goals."
• "Which departments are optimizing locally at the company's expense?"
• "Help me prepare the strategy communication for next quarter's planning cycle."

Run the automated OKR checker (optional)

The skill includes scripts/alignment_checker.py, a Python 3 script that reads a JSON file of company and team OKRs and reports orphan goals, conflicts and coverage gaps. The skill's quick start command runs it from the skill folder. You need Python 3 installed, and you must prepare the JSON file yourself. The source does not define the JSON schema, so check the script itself for the expected format.

The checker only compares goal text and structure in the file it is given. It does not know your business, so treat its output as a prompt for review rather than a verdict.

Example: a short diagnosis workflow

A typical session starts with the articulation test. Ask five people from five teams what the company's most important priority is. If all five give the same answer, the strategy is clear enough to cascade. If fewer than three agree, the skill advises fixing the strategy before the cascade.

Next, the agent maps company OKRs to team OKRs. Any team goal with no parent is an orphan, and any company OKR with no owning team is a coverage gap. Conflicts are goals that, when both succeed, produce a worse outcome for the company. The classic example in the source is sales closing volume contracts with poor-fit customers while customer success is measured on satisfaction scores.

What does the skill need, and what are the safety notes?

The skill requires Python 3 only if you want the automated check. Without it, the agent can still run the articulation test, cascade mapping and realignment discussion from your descriptions. The skill does not use the network, but it does run a script, so read the script before running it on your machine.

Be careful with sensitive data. The skill is not meant for organizations whose OKR data should not be processed through an AI agent or uploaded to a third-party tool without review. Also note that it is not a replacement for real leadership judgment on strategy.

• Requires Python 3 for scripts/alignment_checker.py (optional).
• Requires a JSON file of your OKRs only if you want the automated check.
• Runs scripts: yes. Uses network: no.
• Do not send confidential OKRs through an AI agent or third-party tool without approval.

Tips for getting useful results

Keep the company strategy to one sentence and one page. The skill's playbook recommends a one-page strategy filter that lists top priorities, owners, what the company is not doing, and key metrics. The 'what we are not doing' section matters as much as the priorities.

Run the cascade again each quarter. The playbook describes a quarterly alignment check to prevent drift from returning after fixes.

• Start with the articulation test before judging the cascade.
• Fix incentives before rewriting goals, since rewarded local metrics will override new goal wording.
• Hold a cross-functional OKR review before each quarter begins.
• Run the workshop-style realignment rather than sending a memo.