incoming-request-advisor — Claude Code skill: decode a request
Splits an incoming message into the literal ask and the outcome behind it. Useful when a loaded ping, email or escalation needs a reply.
npx skills add deanpeters/Product-Manager-Skills --skill incoming-request-advisor -a claude-codeWhat incoming-request-advisor does
The skill puts the agent in a product leader's chair and works through an incoming message: a Slack ping, an email, a mandate, an escalation. It pulls the literal request apart from the job the sender is chasing, then returns a Markdown breakdown.
Its concepts are specific. The sender's role is read against your own work — upstream, peer or downstream — because the same words call for a different answer depending on who wrote them. Success criteria (how the result will be judged) stay separate from must-haves (what goes into the deliverable). Guesses are labelled as inferences and collected in a list of assumptions to validate.
The depth matches the message. A one-line ping does not get all twelve sections; empty ones are collapsed or marked as having nothing stated. Each bullet stays short and scannable, like a sticky note. When the sender or the situation is unclear and it changes the read, the skill asks at most three questions, one at a time, then continues on labelled assumptions. The message can arrive as text, a screenshot, an image or a file.
What it helps with
- tell the literal ask apart from the outcome behind it
- read the sender's role and stake before replying
- keep success criteria separate from must-haves
- handle a frustrated colleague's escalation without firing back
- catch a vague feature request that is not a build order yet
Example prompts for the agent
- Break down this message before I reply:
My VP DM'd me: "Any chance the dashboard redesign lands next sprint? Board's asking." - A product lead sent a long escalation email and I need the job-to-be-done first: [paste the text]
- A stakeholder sent a mandate that reads like a build order. Find the outcome underneath it and the assumptions I should check.
How to install incoming-request-advisor
1Skills CLI (Vercel)
npx skills add deanpeters/Product-Manager-Skills --skill incoming-request-advisor -a claude-code2Manually: put the files in place
~/.claude/skills/incoming-request-advisor/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/incoming-request-advisor/
1Skills CLI (Vercel)
npx skills add deanpeters/Product-Manager-Skills --skill incoming-request-advisor -a codex2Manually: put the files in place
.agents/skills/incoming-request-advisor/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add deanpeters/Product-Manager-Skills --skill incoming-request-advisor -a cursor2Manually: put the files in place
.cursor/skills/incoming-request-advisor/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install deanpeters/Product-Manager-Skills incoming-request-advisor2Manually: put the files in place
.github/skills/incoming-request-advisor/SKILL.md1Skills CLI (Vercel)
npx skills add deanpeters/Product-Manager-Skills --skill incoming-request-advisor -a gemini-cli2Platform CLI
gemini skills install https://github.com/deanpeters/Product-Manager-Skills --path skills/incoming-request-advisorManually: put the files in place
.devin/skills/incoming-request-advisor/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add deanpeters/Product-Manager-Skills --skill incoming-request-advisor -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about incoming-request-advisor
How do I install incoming-request-advisor in Claude Code?
Run npx skills add deanpeters/Product-Manager-Skills --skill incoming-request-advisor -a claude-code. The skill file then sits at ~/.claude/skills/incoming-request-advisor/SKILL.md.
What does incoming-request-advisor do?
It turns an incoming message into a structured breakdown with the literal ask in one place and the job-to-be-done in another. It also assesses the sender's position and lists the inferences worth checking.
What do I need to run it?
Just the message, pasted as text or given as a screenshot, image or file. If the sender or the situation is unclear, the skill asks up to three questions one at a time. It follows the workshop-facilitation protocol from the same repository.